mlp-mx6 wrote:What does not make sense to me is to put the one kind of component you KNOW will wear out (the PS filter caps) *underneath* the board to which they are supplying power. You then have to disconnect a LOT of things to replace them. I know it is years between replacements (in a normal scenario) but why make it that much harder? Makes no sense to me at all.
Yeah, I try to avoid that scenario since I like to make one long board. I'm sure it's easier to swap hidden caps on a separate PS board as used in most TWs.
FWIW, I've also used can caps with no probs. My personal preference is to use them for the plates, screens and sometimes PI. They free up chassis space and allow me to distribute individual caps for the preamp. Putting everything on one board certainly makes for a LOOONG peice, but saves me $$ by letting me burn up surplus caps. Needless to say, repairing any wiring mishaps under an installed board is a real PIA!
You do not have the required permissions to view the files attached to this post.
Do you remember this amp. It was supposed to be an octal rocket based amp akin to your ventures but with distributed capacitance in the preamp and a capcan taking up the main power duties.
FWIW, the amp is dead quite
Fishy,
I do remember that amp but I had forgotten about your cap mounting technique. That execution is extremely well done and yet so straight forward. The fact that it is dead quiet along with simplicity doubles the coolness... overall this build is one of those "I wish I thought of that" events.
For those not familiar with this build, it is the amp for the guy who wants everything... well at least more options in output tubes. I was in one of those extended distraction periods when the original post on this came out but this amp is a classic demonstration of cool ideas. The power supply is novel, functional and dead quiet along with the great output tube options.
I will indulge my friendship with Fishy and post his photo here and give you a link to the thread he started on it. Great stuff!
Cliff Schecht wrote:I like the silicone to chassis cap stacking method. It's clever and it's proven to work well. It makes for some long wire runs from the capacitor to the B+ line but they're not that long so everything stays happy (enough). Who knows, maybe the long runs have an intentional affect on the sound..
I have to agree, it has proven itself over time. With your strong engineering background it's interesting and slightly refreshing to hear your mention the potential for an intentional affect regarding the wire length/locations.
Cliff Schecht wrote:I like the idea of shoving the caps below the preamp board too. I could see this working out well if you planned for it (as opposed to retrofitting an existing amp) and you could place those caps where they should be, right on the resistors they decouple.
I've had iterations on this theme that I've implemented with good success. I don't put them below the board but rather on top of it I think I saw it first on the Hoffman builds that Skippy mentioned.
Attached is a photo of the sub-miniature tube Rocket builds I put together. I used the perf board on the first one then moved to a turret board and eventually a PCB board as a production model. All of these amps are really quiet although the total output is also pretty low. I am working with some other guys on the forum to bring this idea to the bigger Wreck builds.
rj
You do not have the required permissions to view the files attached to this post.
I'd love to hear a comparison of an amp, keeping all other things equal, with the decoupling caps distanced away and then placed right on the preamp board. I suspect the difference is negligible in the Rocket but may factor into the uber-sensitive Express.
I've posted my goofy method for getting those electros right up on the resistors. It works especially well when you PTP the lower voltage stuff and turret-mount all of the high voltage stuff. What I end up with (aside from a tangled mess) is the B+ on one side of the turret board and ground on the other. I guess it's hard to visualize without a picture, so here's the 5F6A I threw together this way.. Both builds that I've done this way are very quiet though. The obvious disadvantage is that you have ground and plate voltages sitting within about 1/2" proximity of each other and the builds tend to be "layered" (i.e. I have to remove half a coupling cap to get to some of the V1 stuff.
You do not have the required permissions to view the files attached to this post.
Thanks for the kind words RJ. You were the catalyst for me on this one.
Fortunately, no one had tried the reverb at that point otherwise I might still be trying to figure that out.....
Lots of very interesting and impressive builds in this thread!
In terms of the distance of the filter caps, look at Neumann tube mics. They might be 100 feet away from the power supplies connected by relatively low gauge wire. Not many complaints on the sound of those old mics. In fact I think a lot of high end audio equipment tends to put the power supply in an entirely separate chassis.
I can understand the shorter connections are less likely to pick up (or perhaps radiate) noise, but is there any other advantage? I can't imagine the added resistance of the wire being significant.
Ken Moon wrote:I'm glad I bought the huge size Unibit step drill - I think the biggest headache with using the cans will be making the holes
A spade bit also works very well in aluminum. Set your drill press to the slowest speed, and lubricate the bit well.
For aluminum I was always told to lubricate with soapy water. You keep pouring on the water while you drill the hole. It helps to keep the hole clear of debris too.
I've found that with Allen's chassis lubrication isn't necessary though. If you use a hand drill and a big fat Unibit, you just start slow for the starter hole (getting the long curly aluminum shavings is ideal) and work up the speed as the hole gets bigger. The chassis is thick enough that it won't bend as you start cranking away. Thinner aluminum chassis ala Bud chassis aren't as tolerant to essentially forcing the issue though. Those are where you want to use the slow and steady method (drill press and lubricant).
Haha, yes the soapy water trick! I've drilled thousands of 1"+ holes in thick structural steel with just a Hogan drill and soapy water. Pop a hole in the cap of a 20oz plastic bottle and fill with pink soap and water. Voila! I once had to send an apprentice packing when I caught him filling a bottle with Fast Orange.
Ken Moon wrote:.... I'm glad I bought the huge size Unibit step drill - I think the biggest headache with using the cans will be making the holes
For those that don't know this, a large step bit will make a really smooth and large hole with very little effort on your part, as long as you have the drill strong enough to turn it. I prefer the drill press. In comparison with a chassis punch I find it to be about 90% easier to work with.
Any time you cut a hole a little cutting fluid will do wonders. Soapy water or professional cutting fluid might be the answer. I have also been known to put a little rubbing alcohol on as a cutting fluid. While this is not a good practice for health and safety reasons it really produces great results and then evaporates away with no mess.
Here is a link to photo that Zippy had pointed out to me a long time ago. As I mentioned before this was a bit of an inspiration to me in my board designs.
I don't believe that getting the long curly shavings is ideal.
That usually means the cutting edge is dragging or scraping which means *heat*.
You always want chips. Chips (generally) means you're cutting the aluminum instead of scraping it. And chips means the cutting tool and the aluminum have a chance to cool in between each chip being cut and thrown away from the work piece.
That being said it's really hard to get that with a Uni-bit because it's simply not sharp enough. So lubrication is good. I prefer WD-40 instead of water.
Just my $0.02. Okay back to original thread.
I've tried cap cans, stacks, stacks on board, and on the main board near each stage.
Cans definitely is the cleanest way to go for an almost un-discernable noise difference versus the other two methods. IMHO.
261]
