Hypex UcD Car Amp
I bought the modules here from way back May 2013 and been confused as to which route to take if I would DIY the powersupply, chassis, control boards etc. Fast forward to a few days ago, it dawned on me, why not just chop up a car amp and use parts of it to build a new one. This is what happened.
A DIY chassis would be nice to have a perfect way to showcase the innards if I use an acrylic top cover (as with my previous car amps) Going the chopped up amp route, I would have minimal work with the chassis and powersupply. The result would be a sleeper amp. With one look, you would assume it to be an off the shelf cheapie amp.
I started with an amp from the 90's which I got from a local repair shop. Tested the powersupply (which worked) and a quick look shows that all four UcD180HG modules would fit with minimal room to spare. Like it was meant to be. I was torn on tearing this amp apart as it was a working unit (but a little dirty cosmetically). Thinking I won't be using it in any install then I would just give it a new life instead.
The powersupply already had integrated circuitry for DC protection, overcurrent trip, thermal protection and a delayed mute signal which I can definitely use for the enable pin of the UcD modules. Perfect!
Here is the amp as I got it.
Oh it fits. Barely enough space all around.
The teardown begins.
Analyzing the circuit on how hard it will be to graft the control lines to the UcD modules.
Point of no return. Lots of thinking went into this on where to cut.
Powersupply section separated from amplifier section. The crossovers are on a separate board.
Removed the parts that will be upgraded for higher output since originally,
the amp was probably around 65W/ch. Now it will be 180w/ch.
Soaked the heatsink in detergent and water and a lot of elbow grease later, this is what we get.
Needs a little more work as the heatsink fin edges have scratches.
Powersupply prepped. The old mosfets (IRFZ30) are replaced with IRF3205s,
gate drivers and gate resistors changed to better drive the heavier mosfets.
Transformer (not shown) will be rewound for more copper and higher output voltage.
Test fit in the chassis.
View from another angle.
Gate drive transistors originally used C1815 and A1015 100mA rated transistors.
Replaced with BD139/140 transistors which conveniently has the same pinout.
Elna and Rubycon caps for the main supply lines.
Connectors that will be used. Input will be balanced using mini XLR connectors.
Stay tuned, I'll post updates as this project progresses.
09 Feb 2014:
Some more updates on the build
Modules test fitted.
Transformer rewound with new secondary from DIY litz wire (40 strands of #31 wire).
UcD module bracket.
Wired up the power supply lines.
Used my milling machine to cut out slots for ventilation.
...and the rectangular holes for the terminal blocks.
End panels completed and sanded.
Connectors test fitted.
Russian copper for the internal speaker wires.
First light. No fireworks!
Rats nest complete.
Added ferrite beads on the supply lines for EMI suppression.
Added this transistor to invert the mute signal and make an open collector output
for the enable control of the UcD modules.
More Russian copper.
Design fail. Mounting screw holes ended up under the terminal blocks.
Solved by drilling two more holes.
Not that it needed it, but just did it the same on the other side.
Luckily,mounting holes lined up in the space between the mini XLR jacks.
Canare Star Quad cables, Neutrik XLR plugs and Switchcraft mini XLR plugs
to connect the amp to the DCX2496.
I had a big problem with a very loud turn on pop.
It was caused by the amp turning on before the DCX fully boots up.
I replaced this resistor from 47K to 330K to increase turn on delay to 3.6sec. Problem solved.
And the gain was too high so I removed this resistor to reduce the UcD gain from 21dB to 13dB. Still high but more manageable for my purpose.
Here's the PDF for more info on adjusting UcD module gain.
Current consumption of the amp at the bench.
Supply rails are about 43V at 12.6V in, 45V at 13.2V in, and 50V at 14.6V in.
I was curious to see how the supply rails rise and fall during turn on and off so I used the scope. Seems they charge and discharge at almost the same rate so I think it's good. On the left image, cursor A shows a small overshoot of the supply rail (since this is a semi regulated SMPS) and cursor B is the point where the UcD modules are enabled.
15 Feb 2014:
Gain is still too high for my liking so I installed a passive attenuator using 10K 0.1% 25ppm resistors and Bourns 3362 trimmers.
Prepared the parts.
Trimmers mounted. The threaded standoff is riveted to the PCB so there's
no need for a nut on the other side.
Measure twice, drill once. And it still came out misaligned!!!
Still useable though.
Trimmers connected to the resistors via twisted wire.
Precision resistors are mounted directly to the pins of the mini XLR connectors.
Adjusting the gain to make it similar to the old amps so that there will be minimal retuning.
Amp powered on the bench. Trimmer adjustment screws are a little off center but isn't too bad.
Page created and copyright R.Quan © 02 Feb 2014.