First Watt Amplifier & PSU Build

Hifi2000 drilled and cut the holes in the chassis back panel from the file I supplied.



I was glad to see the power switch fit into place just right, as I found the manufacturers datasheet quite wishy-washy about the size of the cutout and ended up having to try measuring it with a ruler instead. The switch is positioned quite high on the back so it will hopefully be easy to find when blindly reaching to turn the amp on/off.

On the other side more wire prep.



For once I'm using my country's AC colour coding.



There was room enough for the TIO board input tabs to be used at least. Also visible is the transformer ground wire, crimped with a ring, which'll be attached to the chassis.



And this is where I came in.

In an earlier iteration I had made an accommodation for a front panel switch, but sacrificed it for shorter, tidier internal wiring. It likely would have also meant crossing the mains AC and the DC lines - something I really wanted to avoid.

At the moment I'm waiting for an order from Front Panel Express. Along with the obligatory front panel, I also used their software to design the mounting gasket for the capacitors plus some vented top plates for the inductor shrouds. Unfortunately USPS never scanned the package when they took receipt of it, and I have no idea where it is, or if it's traveling through their network.

 

I have a DIY Aleph J that I purchased from the original builder. I also have a real First Watt M2.

I did want to jump in here to comment on a statement made by Nelson Pass with regard to his First Watt offerings.

One of the things that he did was to use the same case design for his different First Watt amps.

Because of the case design and the heat sinking area that was available, he decided to keep the output power of these amps at around 25-Watts.

Nice work and a really nice thread!

 

That's some interesting info on the case S&G, thanks for the message. I think some of Pass' designs can be made more powerful somehow - by chucking more transistors at them or something - but it would make sense a larger chassis size would be needed.

Do you have any comment on the M2 vs the Aleph J?

 

I bought the Aleph J first and thought it was a nice sounding amplifier.

I had an opportunity to buy an M2, so I did.

I don't think that they are too far apart on the way that they sound but I think the M2 was a bit more musical, where the Aleph J was a bit more detailed.

I used them only on my Altec Lansing A7 speakers.

I would consider either of them to be kind of hybrid sounding amplifiers. Kind of like 2/3 solid state and 1/3 tube like.

They basically exhibit the sound signature of SS amps in their clarity and detail but without the edge that is often associated with SS amps.

The Altec A7's are high sensitivity horn speakers with a sensitivity of around 100-dB. This makes them very easy to drive. I often power them with a 3.9 Watt Decware Mini Torii tube amplifier.

So power was not a concern. But I did find them a bit thin, rather than big and full sounding. This is for reasons that I fail to comprehend.

 

That's not the first time I've heard the First Watt amps described this way (although if you've relayed similar thoughts in other threads, I may be recalling that). I've found the F6 to sit in the space between the rich soup of my Pathos Logos and the nimbleness of the Rega Elex-R but the F6 seems a bit dryer than those two, if forced to choose an adjective. I look forward to comparing it to some of the other FW designs.

Thanks for your thoughts.

 

Power Supply Addendum

In the end my design is not so far removed from the original.



This schematic shows that the resistor banks in the filter have been replaced with inductors, and between the transformers and the rectifiers are the additional fuses plus the snubber. Also, the location of the bleeder resistor and inrush current limiter thermistor have changed, but this was only due to reasons of physical placement within the design.

I haven't really explained what the snubber does - so since nothing's happening on the construction front...

I've been snubbed!

The diodes that make up the bridge rectifiers are turning on & off as the AC cycle moves between negative and positive voltage. When the diodes turn off this causes the transformer to 'ring'.

Usually I'd consider it bad form to be linking directly to another forum, but I hope it can be forgiven here. The 'Quasimodo' is a circuit designed to simulate this effect, enabling the derivation of suitable snubber values for a particular transformer without the need for hardcore maths. Simple, no-math transformer snubber using Quasimodo test-jig - diyAudio

Using the Cheapomodo version of the above circuit I tested my transformer. The green line shows the ringing observed with a USB oscilloscope.



While watching the oscilloscope one turns the potentiometer, altering the resistance for the snubber and damping the ringing.



And here is the final waveform. By measuring the resistance of the potentiometer with a multimeter the value of the snubber's resistor is obtained. However, while being an interesting exercise, it's appears this snubber may have no real-world effect on the power supplys performance.

This article by Rod Elliot Snubbers For PSUs , while not naming the Quasimodo directly, is obviously a repudiation of not only the usefulness of such snubbers but also of the Quasimodo circuit itself. Another article addressing the subject more broadly, which demonstrates no benefit to the resulting DC output or amplifier performance can be read here: Taming the LM3886: Rectification and Snubbers

As you can see I left it in anyway. The upshot being that while the benefits are questionable, the inclusion of said snubber will cause no harm either. And, as I believe the author of the Quasimodo has said, isn't the fact that it looks so bad on a scope reason enough?

Two poles, with luck.

The bipolar nature of this PSU probably deserves some mention, because it is some dark wizardry.



The two branches extending out from the transformer are both separate, single, positive power supplies with one +23V rail and one ground. But by connecting the ground of one to the positive of the other it becomes a bipolar supply with one positive rail (+23V), one ground (0V, created by that connection) and one negative rail (-23V, which was previously the ground of the second supply).

The layman may wonder why such a configuration doesn't instead produce +48V, +24V, and a ground. I haven't been able to find an explanation yet; maybe I'm not googling it quite right. Perhaps it has something to do with electron flow or the like.

Thankfully I'm sure, this will be the last post exploding with text. To conclude, here below are the other web pages I found useful along the way.

The Linear Power Supply - DIYODE Magazine
Solid State Power Amplifier Supply Part 1
Linear Power Supply Design

Now it's probably time to be getting on with the part of the build I've been dreading the most - drilling and tapping the threads on the chassis heat sinks.

 

I read First Watt amps are as close as you can get to a quality tube amp. I am not saying a vintage amp with whatever tubes installed. It is stated better. I am saying the better tube amps with best tubes installed. Seeing the damn tubes can cost $800 it makes you think twice.

As for the power supply, I would use IC regulated with .05 millivolt AC ripple. I think the design is too high with .25 volt AC ripple.

 

It's 0.0025 volt at the output according to the simulation is it not, but I get what you mean.

If I recall correctly, the reason regulators aren't used in solid state power amps is due to the comparatively large current swings. Basically, tube equipment is high voltage and low current, solid state is the opposite right?. Do you know of any notable power amps that are regulated? I haven't looked into it, but it would be interesting to see how it's achieved.

I always get the urge to build a tube power amp, but I think the amount of time I've spent fussing with the small signal tubes in my DAC and preamp is enough Cheers!

 

Pass amps generally have good local power supply rejection in each stage, so are not so sensitive to ripple.

Great looking build, btw!

 

Thank Ripblade, and you make a good point. It's entirely conceivable this upgrade will produce no discernible improvement either through perception or measurement. I try not to think about it.

 

I wouldn't go that far, but I don't think it will be the defining quality of the sound of the amp. Every little bit helps, though.

 

It's been a long absence.

The metal pieces that I ordered from Front Panel Express never did arrive. I've not had an issue with USPS before but on this occasion the package missed being scanned at pickup, or at any stage along the route. It must have been lost down the back of someones couch or something.

I waited about 6 weeks before contacting FPE, since international shipping times have been extremely variable lately. Hell, I recently had a tube order that left Ukraine on a plane that somehow spent three months in the air before landing in my country! When you live in this part of the world waiting patiently is something you get rather practiced at.

Front Panel Express were very good about it, immediately producing a new set of parts for me and dispatching them via FedEx free of charge. I imagine this would have been quite expensive for them. The reason I had requested USPS in the first place was that using FedEx would have effectively doubled the cost of the order, not to mention the cost they've borne producing a second run of metalwork.
I feel a bit bad about it. I imagine their business is suffering like many others at this time. The least I can do is give them two thumbs up here.

They package their orders in a pretty nifty way, tightly vacuum packing the parts to a stiff piece of cardboard making it almost impossible for them to shift around inside the box during transport.

 

The thing I needed most in order to continue building was the capacitor bracket. Early on in the design process my only real goal for this project was BIG CAPS. Big beer-can sized caps with screw terminals like I had seen in other builds.

Once the decision had been made to run with a CLC filter there was no more room for big caps but, without thinking, I purchased screw terminal capacitors anyway which meant having to find a way to mount everything without drilling a million holes in the chassis floor.



I started attaching the hardware to the bracket, awkwardly, as there were a lot of little nuts, screws and washers to fiddle with. This is the underside, where the capacitor clips and bleeder resistors are located.

Bleeder resistors drain the capacitor bank when the power is turned off, which is a particularly useful safety feature when a project is being worked on, since the caps may hold a deadly amount of charge for a long period of time. You don't want them discharging into you while you're messing around. Although not strictly needed otherwise, they do however provide a little voltage regulation to the power supply if used during regular operation.



The capacitors were slide into place and I spent ages carefully positioning them while tightening all the hardware bits. I only hope I haven't overdone the clips. How much pressure can the capacitors take before they start to deform I wonder.



The cap spacing was designed specifically for the brass ground bus bar that can be seen in the photo above, but in practice it was too fat and came too close to the opposing terminal on for my liking.

 

fascinating read....good luck with the build

 

Cheers!

 

Here's the cap bracket from the top. Keeping all the mounting hardware underneath should make for a nice clean look on this side.



There's a terminal block on each end, which are where the amp boards will be able to be connected and disconnected from the power supply.

Something that's slightly disappointing is the colour mismatch between the transformer circuit board and the anodising of the capacitor bracket.



The blue of the PCB leans into the green end of the spectrum, whereas the bracket goes the other way toward the purple end. It's fine, but looks a little naff side by side.

 

The wiring of the capacitor bank proved very time consuming. Several hours over three days measuring & cutting, then crimping & soldering each wire to their connectors, and finishing with heat shrink.

Building the filter circuit started by creating the main capacitor connections with 12 gauge wire.



Here the outer DC rails are combined to form the ground, while the inner connections will become the positive (black) and negative (white) rails for the bipolar supply.

Then 16 gauge was used for all of the other incoming/outgoing wiring.



This capacitor block kind of acts as a central hub for the CLC filter.

The black & white twisted pairs carry the incoming DC from the rectifiers, while the colour matched twisted pairs go to the chokes. The green cables are the ground reference lines, and together with their accompanying black & white wires form the +24 0 -24 volts for each side.



I added a thin smear of thermal goop to the underside of the bleeder resistors and soldered to their tabs the capacitor connection wires.

 

And this is close to its final form.



With limited space around it the whole thing's quite difficult to move into position between the transformer and choke enclosures and get connected up . There's not a lot of wiggle room with the wire lengths either, but the test fitting seemed to go okay.

Viewing it from the top shows how the output voltage wiring pops up through the centre of the bracket. I was concerned that the wire insulation was going to rest against the somewhat sharp edge of that middle opening, however it appears not be an issue.



The main purpose of this design was to minimise, as much as possible, the length of the cable to the amplifier boards while also keeping said wiring away from the signal cabling. In this, I'd say it looks to be rather a success.

These output cables connect at the terminal blocks mentioned earlier. Once the bracket is mounted inside the chassis the blocks should be positioned relatively close to the amp board inputs and only require a short run further to connect.



The last thing was to hard tighten the fixings. Because all this wiring will be sitting obscured underneath I was concerned about one of the capacitor clips loosening and a component perhaps contacting the bottom of the case. Now that it's all together it seems highly unlikely.



As a precaution I've used short hexagonal spacers as fasteners rather then regular screws, which enable the addition of nylon bolts to the ends. These should provide adequate insulation if the worst happens. Hopefully it won't be a problem that the spacers don't go very deep into the holes - only about 5mm. I did also manage to strip of one of the internal cap threads enough that I had to substitute a longer bolt.

Seems I'm doomed to de-thread something each build. I'm not sure how my thin, child-like wrists manage it.

Is there anything I could add to the screw fixings themselves to ensure they don't loosen off over time? Hot glue?

 

Now it's time to start the final assembly of the power supply.

Some time was spent sanding away the anodising from the surface of the chassis to ensure proper contact with the other metal parts within the interior. These also have areas which have been sanded to bare metal where needed.



This is done to ensure each part of the chassis and the internal structural elements are all linked to the main safety earth. If any metal portion, through failure or accident, is electrified then there is an escape path through to ground via the IEC earth pin.

Otherwise that voltage may find its way to ground through the user, which is less preferable.

I found an alternative way of connecting the transformer input to the TIO board. Luckily, I had placed a secondary install position for the line capacitor onto the board and was able use this instead.



The input wires are 18 gauge and just small enough to fit through the holes on the board, though I doubt they could have been squeezed in if the wires had been tinned first. The original tab connectors were heat-shrinked over.

 

Impressive work and great thread

 

I appreciate the message, thanks!

 

With a little wiggling around and a little bit of cursing the filter section was assembled. Modular it might be, but not without some frustratingly tight spaces one has to get ones fingers into.

The capacitor bank was installed first and then the chokes connected to it. In the end I brushed the cap screws with a bit of clear nail varnish, hopefully to provide additional protection from them loosening up.



The choke mounting holes had to be aligned with the holes in their enclosures and the chassis both, then the screws carefully inserted through them with some long pliers. Down the bottom of that image I am preparing to solder the power supply ground wire to a thermistor. Something I should really have done before putting in the chokes. There wasn't much room left to poke around in.

Here the safety earth and the supply ground are both screwed tight to the main ground spur.



Below is a beauty shot from underneath. The evidence of my poor drilling has been largely covered up by the washers & bolts.

 

The last thing was to add some covers to the choke enclosures. They were designed with quite large holes as I wasn't sure what kind of temperature increase, if any, to expect from the inductors when in operation.



And now the power supply is complete. For a while there I was sure this project had gotten away from me and I'd bitten off more than I could chew, but it appears to have come together largely as planned - at least visually.

With the protective plastic removed the Toroidy transformer is particularly beautiful, lop-sided logo not withstanding. I don't know if it's removable. When the front panel is on it won't matter anyway.



Since it functioned properly all those months ago when I tested it in my garage I'm not anticipating any problems when I spark it up tomorrow, cross-fingers. However, it won't be truely known if it's fully functional until employed under load.



------------------------------------------------------

Next up - I can't put it off any longer.

 

The only kind of metalworking I've had to do for anything I've previously built has been the kind-of ham-fisted drilling seen earlier in this thread, which is why I was particularly dreading this portion of the project; drilling and tapping the amp PCB mounting holes into the heat sinks.

The 4U height, 300mm length Hifi2000 chassis I used for the first F6 build can be purchased with the DIYAudio universal thread pattern already drilled, as can the 5U 400mm version. But 300mm was too short and 5U was too tall. I wanted 4U & 400mm, which meant attempting the thread tapping myself.

It took a bit of research to figure out how best to do it, and with what tools. I ended up selecting a 2.5mm cobalt drill bit, a fluteless countersink bit (for de-burring), and spiral flute tap. I found out, thanks to a very helpful YouTube video, that spiral flutes are best for blind holes like I was going to be drilling.

Hand drilling this would have been a disaster, but luckily my neighbour is the manliest of men and he let me use his drill press.



Using the drill press was well worth it. It took 2-3 hours to drill the 28 holes needed and the accuracy was as close as one could hope for in a back yard job. I was able to de-burr by turning the countersink bit with my fingers.

My neighbour also suggested tapping the threads slowly using a small power drill on a low torque setting. I was dubious but it turned out to be quite effective, even if it was difficult to keep the tap fully perpendicular



The end results are pleasing, if imperfect. There was little tolerance for error but hopefully I'm inside the margin enough, even though some of the the spacers when inserted aren't entirely 90 degrees upright. It's enough that it came out better then expected, and I am much relieved.

 

Phenomenal work. Have you considered a plexiglass cover for the top piece? I mean, that is some gorgeous work inside there. Shame to hide it.