[tomb]

Quote:

And a reminder to all building the torpedo!
Ground the ALPS RK27 before soldering on the OTs!
Had a small issue with it as my screw driver wasn't small enough, and had a tough time putting on the ground wire.

We used to say it was grounded through the case and the pot shaft, but with the black anodized finish, that's not possible anymore.

 
Good catch.

 

[heku]

Here is my attempt to reduce hum


It is grounded metal box that covers output transformers. It almost completely removed hum.

 

[UKToecutter]

Heku
You could also ground the two heatsinks.
I did that on mine and I have no perceptible hum.

Andy

 

[heku]

I have grounded those heatsinks. With only heatsinks grounded there was noticeable hum.
When I added that metal box hum was reduced to level where I have to listen closely to notice it.
 

 

[KimLaroux]

You shielded the output transformers but not the power transformer? Isn't most of the noise generated by the power transformer?

 

[dsavitsk]

Interesting. Yes, the noise is generated by the PT, and most of it reenters the circuit not through the OPT, but instead through the circuitry itself. Though apparently some comes in through the OPTs, too. It does seem that a combination of things leads to the best outcome -- the CCS mod, grounding the heatsinks, shielding the OPTs, possibly shielding the PT, etc.

 

[heku]

Today I made box for power transformer and now hum is gone.

I also tried only shielding power transformer and leaving output transformers naked.
It did reduce more hum than shielding only output transformers.
But there was still very small amount of hum left, which disappeared when output transformers were also shielded.

 

[tomb]

Quote:

Today I made box for power transformer and now hum is gone.

I also tried only shielding power transformer and leaving output transformers naked.
It did reduce more hum than shielding only output transformers.
But there was still very small amount of hum left, which disappeared when output transformers were also shielded.

Can you give us the details and dimensions on those boxes?  I may try to see if I can get them fabricated somewhere.

 

[heku]

I just used some leftover sheet metal pieces which I had and welded them together.
Power transformer box I tried to make so that there would be some ventilation. Its inside dimensions are few millimeters bigger than transformer itself and there are small feets which leave gap between circuit board and box.
Output transformer box is just solid box that is in touch to circuit board.
I protected nearby components and insides of boxes with dielectric material.
Glue is used for securing boxes in place.
 
I think you would get more accurate dimensions if you take measurements from your own amp.
 
(If you are really anxious to try you could see if cardboard box and tin foil would work. I haven't myself tried this method.) Now I have tried this and it doesn't work.

 

[tomb]

Quote:

I just used some leftover sheet metal pieces which I had and welded them together.
Power transformer box I tried to make so that there would be some ventilation. Its inside dimensions are few millimeters bigger than transformer itself and there are small feets which leave gap between circuit board and box.
Output transformer box is just solid box that is in touch to circuit board.
I protected nearby components and insides of boxes with dielectric material.
Glue is used for securing boxes in place.
 
I think you would get more accurate dimensions if you take measurements from your own amp.
 
If you are really anxious to try you could see if cardboard box and tin foil would work. I haven't myself tried this method.

Well, the reason I asked is that it looks like what you put over the PT is simply some perforated sheet steel.  I would think some tin snips and a vise would be enough to form the box.  You concerned me when you mention "weld," though.  I'm actually quite surprised that the one over the PT has any effect at all with those perforations.  Welding would also not seem to be necessary - the open slots on the sides after bending can't be worse than all those perforations.  Frankly, it seems strange that the box is actually doing anything with all those holes. 
 
Is it completely touching the sides of the PT?  There's very little clearance between the edges of the PT and the PCB pads for the rectifier leads.  Dsavitsk and I had considered trying this before, but it would be quite dangerous in that area with the proximity of the rectifier leads (especially the high voltage ones!).
 
A couple of final questions - from your perspective, was the ripple audible enough to cause you distraction?  If so, what headphones were you using?

 

[heku]

Power transformer box has those perforations only on top. Box is touching transformer from one side because C1 is so close to transformer. There is dielectric material between transformer and box.
IEC inlet side and two other sides have small gap between box and transformer for air flow.
Here you can see those gaps:

 
And here is picture from one side of the box where you can see gap between box and circuit board:

 
Top side of the box is almost touching case lid, so maybe you could do box with side walls tall enough that they touch case lid and leave top side of the box off.

 

[heku]

It was audible with Sennheiser HD800 in high Z setting. With Denon D2000 I didn't hear any hum.
With music playing I didn't notice it but when music paused it was quite noticeable.
 

 

[tomb]

Quote:

It was audible with Sennheiser HD800 in high Z setting. With Denon D2000 I didn't hear any hum.
With music playing I didn't notice it but when music paused it was quite noticeable.
 

Well, I don't want to minimize your efforts, but I'm not sure a Sennheiser HD800 is an appropriate choice for this amp.  I have even said so to other individuals asking questions about it offline.  One would understand after spending $1500 on a pair of headphones that you would want perfect, absolutely noise-free amplification and would indeed, be listening for just that.  However, the Torpedo is on the forefront as a DIY entry-level, tube-transformer-coupled amplifier.  There is no other amp like it that's completely based on a PCB.  In that regard, I think it competes quite well with something like the Bottlehead Crack.  It's actually more versatile in its primary focus with low-impedance headphones such as Grados.  (I would note that the Denon is quite sensitive at 25 ohms.)  Yet, for the world's best dynamic headphones - the logical choice would be commensurate in quality with source and amplification, i.e, "the world's best."
 
I understand that many people who purchase something like the HD800 may be tapped out in their resources and so resort to something more affordable in the rest of that audio string (source-amp-headphones).  It's also quite understandable that you went to these lengths to remove the last bit of hum and they are appreciated, I'm sure, by everyone here.

 
However, I have some concerns and thoughts - some of which are based on those most recent photographs, others are my own review with the amp, its layout, and all the discussions that Dsavitsk and I have had:
 
My concern is that any use of those steel boxes could be quite dangerous. The one over the PT is especially threatening. Somewhere over 200V exists in PCB pads and solder leads that are within a millimeter or two of the edge of that steel. Hell, if the soldering joint on one of those rectifier leads was a little messy, it could be touching.  (This is especially true if someone didn't even go to the trouble of getting the parts flush to the PCB - Dsavitsk and I have actually seen this.)  Similarly, if the amp takes a little jostling one of those boxes could come loose from those glue joints.  Neither the PCB or a steel box offers the kind of porosity needed for a truly robust glue joint.  I can't in good conscience recommend something like this or even provide it myself and depend on the builder installing it correctly and safely.
 
Further, I'm no expert on shielding gaussian fields, but it doesn't make sense that it would work with all those perforations.  Maybe I'm wrong in that regard, but Dsavitsk just repeated that the noise is also picked up in the circuit on the PCB. He's done extensive scope tests in connsultation with several knowledgable people.  In any event, not even a perfectly-shielded box is going to prevent injection of the noise into the circuit on the PCB. The Zener diode mod, on the other hand, removed the last bit of noise to my satisfaction by virtue of its increasing the PSRR to the tubes and output. 
 
I listen to my Torpedo every day and enjoy it immensely (along with the pup

) - it's my primary amp with my Grado HF-1's and 2's, and also my HD600's. We honestly thought that we had removed the noise to the extent that some people might barely notice it, but others would not notice it all.  The Zener diode mod was what led me to conclude that things were pretty good.  Some of you may have noticed after that I black-anodized and laser-etched all the cases I had on hand. I even started offering full kits. Maybe it was dumb on my part, but I have a bunch of new cases and transformers ordered and being manufactured as I type this. I wouldn't have done all this in the least if I thought we had anything more than a triviality.
 
It may very well be that there are some headphones that are not really appropriate for this amp.  The HD800 may be one of those.  Just my own opinion (and beating a dead horse, I suppose), but I would never consider applying a $300 amp with $2 tubes to an HD800 and expect perfection.  However, heko has provided what appears to be a fix to his satisfaction and his work is commendable.  If someone else wants to go that route with an HD800, that's their choice, but I don't think I would personally recommend it. 

 

[KimLaroux]

Tomb, this is DIY, your kits don't have to include all the possible modifications. 

 
If you judge your kit to perform to an acceptable level, then you can pat yourself on the back and take a break. For me that's a success. You can then concentrate on distribution.
 
But if builders find tweaks that presents improvements in their configurations (read - with their headphones), then let them share with the community. If others want to try them, let it be at their own risks. Tweaks don't always work for everybody, or are even necessary to everybody. But it's great to know, from a DIY'er standpoint, that the kit is open to all sorts of modifications with potentials improvements. Isn't this what's so exciting about DIY?
 
I do agree with you that those metal boxes present electrical shorts hazards. Especially the PT one sitting over the rectifiers with nothing more than hot glue... lol. Actually when I asked about not shielding the PT in my last post, I expected those rectifiers to be part of the reasons not to shield it.
 
As for perforations: If the hum is a low frequency, like AC hum, then it'll get filtered out my the box even if it has holes in it. Think Faraday cages. The higher the frequency of the radiation, the smaller the holes have to be. Wifi and cellphones RF needs solid ferrous metal to be shielded off. In my amp, I used iron cages with a hole in the top to shield the tubes from magnetic coupling from the PT. It works perfectly, even with this huge hole on the top.
 
Another, safer approach to shielding in the Torpedo may be to mount metal plates to the top cover. Though it'll mean drilling holes in it, so anyone who'll want to experiment with this should get a spare top cover just for tests. It's something to try, anyways.

 

[tomb]

Kim,
 
Part of my response has to do with several strings of parallel PM's that I had going on at the same time.  Sometimes my honesty is in direct conflict with my self-interests.  It gets frustrating with people asking me to recommend an amp and by the time I'm through with them, Beezar's amps are no longer considered.  It happens a lot.  Maybe it's all these $1500 headphones out there and I refuse to say a $100, $200, or $300 amp is the finest thing available (unlike some other people).

 
I certainly agree with the DIY aspect and would never, ever restrict anyone's experimentation, except when it comes to a safety issue - and I think this may be one of those.  It's probably a moot point, since not many people are going to go to the effort to weld a steel box.  Still, you would be shocked (pardon the pun) to see what some people will do, even in the face of killer voltage.  I'm not saying at all that's what heku has done.  Indeed, heku's work is greatly appreciated and I stated so - plainly.  Others may not have the skills involved, though, and the Torpedo is supposed to be an entry-level, high-voltage amp suitable for someone who's never messed with high-voltage before.  Heck, it's the first and only high-voltage amp I've ever built.
 
Meanwhile, we have and will continue to tweak the amp and if there is mod worth considering with beneficial results, it'll be incorporated and documented as quickly as possible.