• Please note that Audiokarma will be offline briefly for updates early on Monday morning, September 28th.

Steampunk Amp Thread

soundmotor

super modified
Here's my contribution (pix first, details later, crappy cameraphone regrets) -

PIC-0266.jpg


PIC-0248.jpg


PIC-0251.jpg


PIC-0253.jpg


PIC-0257.jpg


PIC-0260.jpg


PIC-0265.jpg


PIC-0263.jpg


PIC-0245.jpg


PIC-0267.jpg


PIC-0270.jpg
 
Register to hide this ad
Details

I found the case at my town dump. It was originally a watch demagnetizer from a pre-WW2 jeweler or thereabouts. The top had a 2, 2"x5" slots in it each with a hardwood insert. The guts were long gone from it so I don't know what exactly was in it or how it worked. It was just the shell and top & bottom covers. This is the tag that was on it -

PIC-0269.jpg


I cut the center divider between the two open sections in the top away to make a single large opening and then used a router table to square it up. Both the top & bottom are sandcast aluminum and the center section is sheetmetal. I painted the top & bottom with Krylon wrinkle paint and once all the holes are cut for binding posts and input terminals I'll paint it the same way.

The screen is stainless steel scrap that I picked up somewhere and the FH allens are also stainless and countersinked into the top. As the top is sandcast, only the presenting surface is fairly square. The underside is rough & knobby so I used small brown bakelite stand-offs to hold the screen off the surface evenly.

PIC-0264.jpg


The large hole in the front perfectly fit a large machine age lamp indicator I've been wanting to use but the vintage of it didn't look right on anything I'd tried. I think it looks perfect on this. Instead of a lamp I will use a superbright LED in place of the bulb. The on/off toggle switch is a DPDT which I'll parallel & capacitor bypass for long life. I still need to find one of those aluminum flags that goes over the switch threads that says "ON-OFF".

Lastly I need to decide what amp will go into the case. I've got a few candidates -

1) I've got a Technics SA-80 receiver. It uses an STK stereo output stage for 15WPC and the circuit is straight out of the Sanyo data sheet at first glance. I may just saw the PCB up at the output stage and hardwire new inputs into the lands and move the power supply & transformer over to the new case along with it. The STK heatsink will fit inside the case with room to spare. As an aside, if anyone has a schematic for this receiver, I'd love to get a copy.

2) Another STK based unit is from ADS. One of the engineer's gave me an amp he built from a P80 Power Plate car amp that had a blown power supply. This amp was good for 40W into 4-ohms on each channel. He made a new supply from an Atelier A2 integrated amp and packaged the whole thing in a case that looks like 1/2 of a PA4 except not as deep. (I am iffy on ripping this one apart and may just rebuild the chassis & square things up.)

3) I have a dead HTiB unit that used TDA7293's for each channel output. They are essentially ST's answer to National's LM3886 "gainclone" and each is good for ~75W into 4-ohms. The unit I have has 5-channels on a single PCB. I figure with some judicious scalpel work I can free two sections off the main PCB for use in this amp. I'll move the power supply over as is.

4) I have a full-range class-D stereo demo board that is good for 75WPC. It uses a compact computer switching supply and the amp itself if about the footprint of a pack of smokes. This one will be the least work and the one I am leaning towards the most. It is also the one most diametrically opposed to the case styling and era.

Comments on which amp to use will be appreciated. I did discount a tube amp for this out of hand. I've got the parts but want to get this completed in a reasonable amount of time. Tube amps always take me a long time when I build them from scratch.

:arrow:
 
I love it, a man after my own heart, reuse, recycle and create something completely different.
 
I'd go class D - biggest contrast between old / new. Unless you have some globe 10s or 50s for the output stage...
 
I'd go class D - biggest contrast between old / new. Unless you have some globe 10s or 50s for the output stage...

Yeah, it's down to the class-D demo board or the SA-80 goes under the knife.

I'll do a tube amp along these lines later but it will use either an 829 or this -

gl815.jpg
 
I'd go class D - biggest contrast between old / new.

Running Update -

I found something else, something I forgot I had. It is, a Tripath TA2020 (class-T) demo board and that is what is going into the enclosure. It is far less power than the class-D one (20W vs. 75W / channel) but that should be enough as I will be using this amp in conjuction with a powered subwoofer. The LM4780 gainclone came in a close second and I may ultimately replace the TA2020 with that later. However, since I have to build the gainclone from scratch, it will add time to the project. The TA2020 I have right now, and it has been running for the last couple days without issue. In other words I can concentrate on mechanical fabrication rather than electrical circuit proofing and hopefully be using the amp by the end of the month.

The demo PCB I have is actually a bit odd. There are (2) TA2020 chips on one PCB. They are stereo chips but the PCB only has one set of stereo outputs not two sets. I have not traced the chips back but I think they might be paralleled up for higher current. If so, it would be a pretty neat piece. In any case it really peforms nicely and should be fine for my application. Sound quality is on par with a Sonic Impact SI-5066, albeit one on steroids.

I'll post a picture of the TA2020 demo board tomorrow.
 
The TA2020 are probably bridged for one mono amp per chip. Gives more power with a given supply voltage, effectively doubling the possible voltage swing across the speaker.
 
The TA2020 are probably bridged for one mono amp per chip. Gives more power with a given supply voltage, effectively doubling the possible voltage swing across the speaker.

I thought the same thing but after reading the datasheet came away thinking not. The datasheet says they already have bridged outputs and if you look at the block diagram on page 5, it looks like they certainly do -

TA2020 Datasheet

It does not call out that you can parallel them but since it is brickwall filtered and the speaker is not part of the feedback loop I am not sure why you couldn't? Any other thoughts?
 
Here is the demo PCB I am going to use. I am taking off the onboard RCA's and adding coax to the ones that will be on the chassis. The output wiring however is of decent quality and unless it is not long enough, I will leave it as is. I am also going to ditch the electrolytics in the input signal path.

ta2020.jpg
 
Some other odds & ends. The heatsink is solid copper & ~0.5" x 2.5" x 6". Overkill for Class-T but I have it and intended to used it with the LM4780. I've got a couple others to look at as well. If I use it I'll cut it in half. Also shown is the AC input filter and a wicked, Bulgin, English-made fuseholder that I have been sitting on for ages. Nickel-plated, bakelite, & phenolic, it is perfect for this amp.

heatsink.jpg


(Eventually I will high higher resolution photos but my Blackjack is all I have today.)
 
I thought the same thing but after reading the datasheet came away thinking not. The datasheet says they already have bridged outputs and if you look at the block diagram on page 5, it looks like they certainly do -

TA2020 Datasheet

Whoooops, right you are. Guess I should have looked at the datasheet instead of ASSuming:D I built a board/amp with the TK2150 chipset, and it had single-ended outputs. Too bad the Tripath business kinda went in the crapper, they had nice stuff.

It does not call out that you can parallel them but since it is brickwall filtered and the speaker is not part of the feedback loop I am not sure why you couldn't? Any other thoughts?

Curious, I guess just look at the board traces to see what's goin on.
 
Very cool! I would love to see some tubes in that case. :D

I could do an SE EL84 in there but I don't have any decent iron right now. The case seems roomy until you start loading stuff into it and I don't want an external power supply. For me at least, scratchbuilding a tube amp is always a lesson in trial and error even though I think I do a pretty good job of thinking through the details. Inevitably I end up moving things...and redrilliing things. Because of the uniqueness of this case, I only want to drill it minimally and not have a bunch of oops holes.

As it is, I am going to build mounting plates for the right & left inside sides. I have some G10 fiberglass sheet for that. The amplifier board, heatsinks, power supply, input filtering, and so on, will get mounted to the plates then the plates will get mounted with standardized hole spacing to the insides of the case. That way if I make a change later such as go to a gainclone instead of the Class-T, I won't be Swiss-cheesing the case sides with new mounting holes. The inner plates can be remade and are sacrificial but the outside of the case will show the same allen head screw positions.

Tubes will likely be in the signal chain though. I am going to build a preamp to go with this and that is where the tubes will reside. Haven't decided on the topology yet but I do however have a VU-meter, possibly the coolest one I've ever seen. ~2"x2", sealed, WW2 military spec. It is just a ridiculously neat item. I've got a pile of Tele ECC88 diamond bottoms and there are some sweet, low voltage preamp designs that I am curious about.

G10 -

g10.jpg
 
Whoooops, right you are. Guess I should have looked at the datasheet instead of ASSuming:D I built a board/amp with the TK2150 chipset, and it had single-ended outputs. Too bad the Tripath business kinda went in the crapper, they had nice stuff.

Curious, I guess just look at the board traces to see what's goin on.

Your guess was mine too. I hate tracing solid-state chips out but am very curious what is up with it. Yesterday while I ran it, I tracked temperature with a laser thermometer. When no signal was present it seemed to idle at a lower temp and when signal was present it would climb. Both sides tracked fairly parallel in this regard so I believe that both TA2020's are seeing signal and outputting it.

I agree on the Tripath products. They made some great pieces. By all accounts the TA2020 is one of them. Cirrus appears to have curtailed all of their standard designs. Could be they will resurface again after Cirrus loads them into their product roadmap. Hope so as they offered audio enthusiasts a low cost way to get a decent amplifier. I've made some nice trades/buys over the years and ended up with a few Tripath demo boards. This isn't the last one I am going to repurpose and turn into an audio component that is for sure!

;)

Now, onto the power supply! This is leading the way -

Astrodyne
 
Back
Top Bottom