I'd just keep the standoffs near the heatsink but not try to screw then down. You do want the standoffs to take the load off the 3886 pins so maybe a little bit of tape or other thin shim material underneath the standoffs so that the opposite board side standoffs aren't "lifting" the heatsink side standoffs off the chassis.
I definitely would want the 3886 directly mounted to the heatsink no matter what. Rube Goldberg thermal energy transfer is perilous.
Thanks for the info. I understand the need for as short as possible wire from pin1 to the chassis. I also get the impression that no other ground wires should be connected directly to or near it. Is that correct?
Based on your Mod86 diagram I think the ground connection at J3 is the signal ground. It is connected to Power86 J2 GND connection which is connected to J1. The wire leaving J1 is then connected to the chassis. Is that correct so far?
If so this makes me ask if the wires, from J1 and the IEC, can be connected to the chassis at the same point with one screw?
Based on your Mod86 diagram I think the ground connection at J3 is the signal ground. It is connected to Power86 J2 GND connection which is connected to J1. The wire leaving J1 is then connected to the chassis. Is that correct so far?
If so this makes me ask if the wires, from J1 and the IEC, can be connected to the chassis at the same point with one screw?
No kidding. To speed up progress (pronunciation guide: PROUGH-GRESS) I went ahead and ordered the TimeStretcher 2000 from Amazon.
Tom
The connection from pin 1 of the XLR input connector to the chassis should be as short as possible. Whether this connection is shared with other grounds, I don't think matters.
Pin 1 of the signal input connector (J2) on the MOD86 is the signal ground. Pin 2 of J3 of MOD86 is the power ground. The two join at the output connector (J2).
You should have a wire from the ground terminal of the IEC inlet to the chassis. The wire should be as short as possible. From this ground point, route to J1 pin 6 (marked GND) of the PWR86. You'll find the wiring diagram in the design documentation for the Power-86.
Pin 1 is marked in two ways on the board: 1) with a little '1' by pin 1 and 2) by a rectangular pad.
Tom
Last edited:
H.Ott explains why one should never use a drain wire along the side of the shield.
It concentrates the current towards the drain wire and ruins the shield performance.
Yes pin1 via very short low impedance connection direct to Chassis at both ends.
Protective Earth (PE) to chassis near the mains entry location.
Transformer interwinding screen to chassis via a very short low impedance connection direct to chassis.
All three of these connections do NOT need to be connected directly to each other. They perform BETTER by being located where they are neeeded.
To refer to them as grounds just confuses the purpose and the assembly and for Beginners the word "ground" completely confuses them.
Note in my first para I never used the word ground. It is not needed.
Last edited:
Thanks Ian. I also plan to follow Tom's wiring directions. I'm just curious why someone would suggest doing it differently. They must think there is some benefit.
I've noticed that answers, to posted questions, are sometimes contradictory. When I find that is the case I usually ask the designer.
I've noticed that answers, to posted questions, are sometimes contradictory. When I find that is the case I usually ask the designer.
I do believe that those who insist on doing things differently believe there is a benefit to their method. Sometimes there is a benefit. Sometimes the change makes no difference at all and at other times it degrades performance. In some cases the degradation can be severe. The giant spider PCB layouts of star ground is an example of well-meaning people who sorta get the idea but in their attempt to address one issue end up creating another. I measured the impacts of the star ground on the performance of the LM3886 many moons ago (link).
In case of the electrostatic shield, Andrew is right about at least one thing: The wire from the shield to ground should be short (low impedance). Similarly, the ground wire from the Power-86 to the chassis should be short. As long as these two conditions are met, I really don't have an issue returning the transformer shield to the Power-86 board. If you keep the board close to the transformer and rotate the transformer to minimize the wire length we're talking maybe 10-20 nH of additional inductance in the shield connection. Not earth shattering and it makes for easier assembly and fewer support emails.
Yeah. A buddy of mine calls the Internet the Contranet for that very reason.
Tom
Tom,
Contranet nails it in my book. I'm going with your wiring scheme. On an unrelated topic, I'd like to know why should I use the 7721-7PPSG* shoulder washer. Is it because it won't draw heat from the heatsink? If so wouldn't a nylon screw be icing on the cake?
However, I have a lot of #4 stainless steel flat and lock washers and 4-40 screws and was going to use them just to cleanup the clutter on my bench.
Contranet nails it in my book. I'm going with your wiring scheme. On an unrelated topic, I'd like to know why should I use the 7721-7PPSG* shoulder washer. Is it because it won't draw heat from the heatsink? If so wouldn't a nylon screw be icing on the cake?
However, I have a lot of #4 stainless steel flat and lock washers and 4-40 screws and was going to use them just to cleanup the clutter on my bench.
Last edited:
By default the Modulus-86 BOM specifies the insulated version of the LM3886, the LM3886TF. No shoulder washer is needed for it. Just give it a light coat of thermal grease on the back, just enough to fill the voids and gaps in the surface roughness. Then fasten the LM3886TF to the heat sink with a #4 or M3 machine screw. I prefer to use a flat washer between the screw head and the LM3886 as well.
The shoulder washers are used with the un-insulated LM3886T. It has a metal tab that is mounted on the heat sink. The metal tab is connected to VEE. To avoid shorting VEE to the heat sink, you need to isolate the heat sink and the mounting screw from the tab. So you use a thermal pad between the LM3886 metal tab and a shoulder washer on the mounting screw.
Never use a nylon screw for mounting semiconductors to heat sinks. Nylon tends to stretch - in particular with repeated heat-cool cycles.
At safe & sane heat sink temperatures (such as the <60 ºC I specify for external heat sinks), there's no advantage of using the non-insulated LM3886T. Only if you really want to push the heat sink temperature to the max (say 90-100 ºC) do you need the non-insulated package.
Tom
really new to electronics
This question will make it obvious how wet behind the ears I am. I'm using nylon insulated faston female quick connects, for the Mod86 build, and don't own a crimping tool. I do have a 2 ton and 12 ton press but don't think those are appropriate for crimping.
I started searching the web for hand held tools and was shocked at the prices...$$$$. What are people using that is reasonably priced?
This question will make it obvious how wet behind the ears I am. I'm using nylon insulated faston female quick connects, for the Mod86 build, and don't own a crimping tool. I do have a 2 ton and 12 ton press but don't think those are appropriate for crimping.
I started searching the web for hand held tools and was shocked at the prices...$$$$. What are people using that is reasonably priced?
For mounting LM3886 I use these clips MAX08NG Aavid Thermalloy | Mouser Also available from Fisher at Farnell/RS. Those put pressure on the right place of the chip and allow for "thermal play".