Technical Assistance/Advice Thread

[ibuski]

While building DYNAHI PS, I'm connecting the earth directly to the chassis with an exclusive screw as it was recommended by many.  You see thick black cable on the photo, excuse me I don't have green/yellow colored cable that is German rule.  I understand the purpose is for safety.   So the ground level have to be shared with the case surface overall.  However Modushop case components with cool black paint seems not conductive between them.   I had to peel the painting off below the washer on the bottom plate.   Then how can we ensure the electrical connection between parts, side heatsink front panel and top plate?   The supplied threads are also coated black and looks not conductive.   Originally made thread holes are also coated, so using chrome plated screw doesn't help anyway.   When the bottom plate is isolated from others, I see it's not useful to connect the ground wire to the bottom plate.   I know some mass production electrical products have a kind of metal wire net like parts between the external panels to guarantee low resistance between them.   Do you DIYers do so always?   Are you (Should I) so careful about the earth/ground connection between the case panels?   I'm wondering if I should connect DYNAHI PS ground to the chassis, too.

[Pars]

For the case, the anodizing will isolate each panel from one another. Many people use a bit of sandpaper or abrasive to clear the anodizing in a small area in between panels, and where the PE ground is connected from the IEC terminal.

As for the PSU ground, I would not connect it to the chassis directly, but use a ground loop breaker (see https://www.diyaudio.com/community/threads/audio-component-grounding-and-interconnection.163575/).

For the case, once you are satisfied, I would check with an ohmmeter to make sure you have continuity.

Edited December 24, 2021 by Pars
fixed link

[Beefy]

1 hour ago, Pars said:

For the case, the anodizing will isolate each panel from one another. Many people use a bit of sandpaper or abrasive to clear the anodizing in a small area in between panels, and where the PE ground is connected from the IEC terminal.

As for the PSU ground, I would not connect it to the chassis directly, but use a ground loop breaker (see ).

For the case, once you are satisfied, I would check with an ohmmeter to make sure you have continuity.

I like using internal tooth lock washers for the grounding points. The teeth cut through the anodising and paint, and prevent the screws from loosening.

Also, I use more than one chassis grounding point (usually at the connection between the side and base plates), just to be absolutely sure.

Edited December 23, 2021 by Beefy

[ibuski]

Thanks Pars,  Yes, I'll make sure the contact between the housing panels.   As I have a motor tool, peeling coating shouldn't be big work.   I wasn't aware of the necessity of the ground loop breaker.   Unfortunately the link you gave me don't have active link to the Figures, but I think I got the most of contents.   I found the design of ground loop breaker on many sites and I'll apply it in my Dynahi PS.   I read that the amplifier chassis and the power supply chassis should be connected with an exclusive earth line not shared with the ground line of the power supply, right?   I think I can follow this, too.  As the number of wires in the umbilical cable is limited, it's a bit pain to lose one for earth though.

Thanks Beefy for the tips to use internal tooth lock washers.   It's nice to know that the teeth can penetrate the anodising coating.  When it's such easy, applying more than one grounding point should be easier.

For me it was good to confirm that you DIYers respect that grounding rule and spend much effort to guarantee it.  I'll follow your methods.

[luvdunhill]

Do ya’ll think the Dremel (Moto) scroll saw would work okay on 1/4” thick aluminum? The manual says 20ga is recommended thickness… so I am guessing not… I am not sure what might be good for this - I need to cut a curved shape in the end of a piece of aluminum. 

 

[VPI]

[Pars]

I see Jeff provided the proper HC advice?

If you have some scrap 1/4" aluminum, I'd try it with the Dremel. Aluminum is pretty soft, so it might work acceptably? You may go thru some blades, so some spare(s) would be a good idea.

[n_maher]

A milling machine is what would probably be good for cutting 1/4" aluminum.

[Pars]

Transformer mounting: for an Antek 50VA, for mounting to a 3mm aluminum bottom panel, I would like to countersink the mounting bolt. Not sure what size of bolt Antek uses, but Avel Lindberg uses M5 bolts. Could I get away with countersinking an M5 in 3mm aluminum? Or would an M4 be big enough for the transformer (I'm thinking yes, but wanted my daftness evaluated before I do any drilling).

[spritzer]

Antek uses M8 bolts and the 50VA units are pretty heavy so a M4 would be pushing it. 

[Pars]

Transformer is an AS-0532, weight 2 lbs. Maybe an M6? Playing around in Front Panel Designer right now and it will let me countersink a hole for 3mm panel past an M8. I guess it would hold OK if not abused too much?

[luvdunhill]

2 minutes ago, Pars said:

Transformer is an AS-0532, weight 2 lbs. Maybe an M6? Playing around in Front Panel Designer right now and it will let me countersink a hole for 3mm panel past an M8. I guess it would hold OK if not abused too much?

M6 should be fine for that - I would reach for a 1/4” screw on instinct so that checks out.

[Pars]

Thanks Birgir and Marc. My main concern was countersinking something that large in a 3mm thick panel. You guys think that would hold alright? I guess if I have to, I could use a pan head and a washer if necessary. Just trying not to have anything hanging down.

[luvdunhill]

2 minutes ago, Pars said:

Thanks Birgir and Marc. My main concern was countersinking something that large in a 3mm thick panel. You guys think that would hold alright? I guess if I have to, I could use a pan head and a washer if necessary. Just trying not to have anything hanging down.

You could use a low profile socket head bolt / wafer head instead of a typical angled - I think FPE also has presets for a larger angle countersink (which exists for this reason) but make sure you can find the fastener before you order the panel l.

[Pars]

I have the panel(s), just using FPD to layout placement of shit. Slimline case for my Salas phono stage.

[gepardcv]

@Pars: I used an M6, countersunk, to mount an AS-0518 on a Modushop Galaxy 388 base panel, which I believe is thinner than 3mm (supposedly 2mm). Also used FPE and FrontDesign. It turned out fine and looks plenty sturdy to me. I don't think the trafo will fall out.

[Kerry]

You could use something like this Rounded Socket Button Head

Mount it from the bottom with the nut on top. 

Not sure I would go to deep on the inset.  Not more than 1mm. 

[Pars]

Yeah, I have used the button head screws before, and may just do that with no inset.

I was going to place the transformer in the vicinity of a chassis foot, so trying to think of ways that would allow me to place both where I wanted without worrying about interference.

I had also thought of raising the transformer up 5mm or so off the case floor, but coming up with something to do that with had me stumped a bit. The transformer is about 4" in diameter. It is a 2U case so I have plenty of clearance height-wise.

Edited January 18, 2022 by Pars

[gepardcv]

I would appreciate a bit of advice with regard to mounting non-isolated active devices on heatsinks.

While trying to get my feet wet with speaker amps, I bought a Gainclone kit (LM3886-based chipamp). The active devices included all require heatsink mounting, and all are non-isolated (LM3886T in particular instead of the isolated 3886TF). The kit does not include isolation pads, PEEK screws, shoulder washers, or any of the things we use here to attach high-voltage devices. I asked the vendor about this, and he said that I shouldn’t worry about it — he said that the tabs are grounded, and therefore do not need to be isolated from the chassis. I don’t understand this statement, and asked if “grounded” in this case means “the case is connected to safety earth on the inlet”, and if that implies that the amp cannot ever be connected to an outlet without a safety earth line (which I normally think of as necessary to protect against accidental shorts, but not as a hard requirement to prevent electrocution in everyday operation). He replied: “The power supply uses a UL Listed transformer, power entrance, fuse block, and power switch.”

This does not sit well with me, but before I push back on the vendor, I thought I’d ask here if I’m just ignorant. Is there something about UL Listed hardware that prevents a high-voltage device which is electrically common with the amplifier case from shocking the user? Is there a grounding scheme that negates this danger?

Edited February 25, 2022 by gepardcv

[audiostar]

For isolation I would use Aavid insulators, mostly ceramics if there is enough clearance, Aavid PPS shoulder washers and stainless steel mounting hardware tightened to 0.8Nm (M3 or #4-40 allen head screws).
The Aavid insulators and shoulder washers are available in different sizes, depending on transistor packaging and size. 
I am attaching an ONsemi paper as well including mounting considerations for power transistors.

For grounding follow this scheme, which shows pretty much the most complex case of a two box design, amp with a separate PSU chassis. Pic shows single ended amps,  balanced would be slightly different as the signal do not touch ground, the amp chassis would change so that the inverted signal (cold, pin 3) would go to a second amplifier board instead to ground. Overall idea should be pretty clear however.

 

Mounting Considerations - AN1040-D.PDF

Edited April 8, 2022 by audiostar

[audiostar]

On 1/18/2022 at 5:58 AM, Pars said:

I had also thought of raising the transformer up 5mm or so off the case floor, but coming up with something to do that with had me stumped a bit. The transformer is about 4" in diameter. It is a 2U case so I have plenty of clearance height-wise.

Using rounded button head would be classic. If not you might pot the transformer into a transformer case like this and then mount it using 4 smaller screws, which might be then countersink. You can then easily rise the transformer as well.  

[simmconn]

I stumbled across the Electrical Spacing section of the PCB calculator in KiCAD, which is from IPC2221A, Generic Standard on Printed Board Design.

According to IPC2221A, most of our high voltage e-stat amp PCB build will fall into the B4 and A6 case. It looks like we may have a problem with the clearance. The biggest problem is with the IXCP10M90S footprint, where the conductor-conductor spacing is only 0.5mm. The SiC FET footprint is better at 1.7mm. Some minor problems here and there where the spacing between B- and other nets are about 0.5mm. I don't think anyone is conformal-coating the board after assembly. So in order to be fully compliant, especially for the (semi) commercial builders, it would be better to revise the PCB.

[JoaMat]

I came across IPC2221A years ago when I started to make my own uncoated home-made boards falling in category B2.

Following the document would have had a huge impact on my layouts. But as a DIY:er I’m entitled to do whatever I like so I went with a clearance of 40th as standard with ground plane both sides. So far, I haven’t noticed any problem due to the clearance.

[spritzer]

40th is fine for our use unless we go to 600V or higher.  We learned this the hard way back in the KGSSHV days. 

[simmconn]

On 4/8/2022 at 1:38 AM, spritzer said:

40th is fine for our use unless we go to 600V or higher.  We learned this the hard way back in the KGSSHV days. 

 

Thanks for the info! It is definitely helpful when I experiment with higher supply voltage next. I'll probably mill away some copper and apply some high voltage RTVs.