Audio Asylum Thread Printer

DanK tubes

Here is the latest project I have been working on. It's a pair of 417a/45 (or 483) monoblock amps. I thought people here might like to see them since they use a total of 10 MQ transformers (5 for each monoblock). These are heavy amps, but worth their weight in gold.

My design principles included:
- a good power supply
- solid mechanical build
- reducing HF oscillations
- eliminating ground loops

Thanks go out to Gordon Rankin and Thorsten Leosch for their comments on using a 417a as a driver. I think the 417a driving a 45 with a grid choke is a very special combination. Here is the schematic I used for the project.

I bought the chassis from a ebay seller in Taiwan, and used Front Panel Express to make a pair of 4mm thick top plates.

For the power supply, I used a MQ R-480 700VCT transformer. This is a nice quiet transformer with low flux leakage and electrostatic shielding. Probably overkill for these amps, but I like to build my projects bullet proof. The recitifier is a 6BY5 dual damper diode. Initial filtering is done with a small input cap (1.5uF) cLC filter (100uF ASC oil cap here) to get a B+ of ~410VDC. It's not really a choke input but is does cut down on some of the HF spray that a large cap input has. I then use a Swenson/Pimm series regulator to step the voltage down to ~330VDC. This regulator is based on Gary Pimm's CCS design.
http://pimmlabs.com/web/regulators.htm
Don't be afraid of series regulation! As Gary has shown, this regulator gives -120dB rejection from 120Hz at 100mA. That's plenty of slew rate for 45 monoblocks. A low noise power supply with 0.2ohm output impedence gives a really nice backbone for this circuit. Thanks Gary and John!

At the input I'm using the B7 10k:10k Ni transformer for galvanic isolation of the source as well as some common mode rejection. I was worried about magnetic pickup so I mounted these inside of some hammertone steel cases. I lined the inside of the steel cases with mu-metal shielding. In the end they kind of look like old Altec transformers. Attenuator is a 10k PEC mil spec unit. Here are some pics of the cases.

The 417a is quite a good oscillator so I used a ferite filter on the heater lines, cut down the tube socket pins, put 10k carbon comp stoppers for each grid pin, and 100ohm carbon film plate stoppers. Thanks Eli Duttman for those tips! No sign of oscillations. I was forced to use an electrolytic cap for the cathode bypass and did so with a 100uF Elna Silmic. I do think these are the best electolytics that are readily available on the market today. The 417a is cap coupled to the output tubes grid with a BCP-16 M4 grid choke. I used a 0.22uF V-cap TFTF here. Each form of coupling has its own drawbacks. I think that cap coupling with the best cap you can find is a good compromise.

The R-480 transformer has a 5V line that I used for the output tube filament. I hooked up each end of the 5V line with a 1ohm/3W dropping resistor to get 2.5V on the filament, as done by Kurt Strain in his Cobalt 45 amp. I then connected a switch across the resistors to short them out when needed. That way I can run either a 45 or the 5 volt equivolent 183/483 triode. With the 183 I can get a little more power since the bias point jumps from ~54V on a 45 to ~58V on the 183. I still haven't decided which tube I like best. I used a 30uF MGBO PIO cap for cathode bypass duty. AC on the filaments, with a 10 turn wire wound hum pot.

The output transformers came from one of Mike's special runs. They are Ni pin stripe RH-PF mounted up with bell ends. The plate choke is a BCP-15 gapped for 40mA. I used another Russian cap for parafeed duty, 4uF 600V KBG PIO.

Thanks go out to Dave Davenport for his treatise on grounding.
http://www.raleighaudio.com/Audio Component Grounding and Interconnection.pdf
I used star grounding with Dave's suggested ground loop breaker. I used Doc B/Paul Joppa's transformer grounding scheme. All transformers are insulated from the chassis with nylon shoulder washers. I attached a star washer/solder lug to one of the mounting screws on each transfomer. Then soldered wires so that each trasformer is grounded at a single point. No sign of ground loops anywhere.

Resistors are a mix of Ni wire wound, Takman carbon film, and AB carbon comp. Soild core copper wiring. etc. etc. And all components are mechanically fixed in place.

For all the bench racers, here are some measurements taken across a pure resistive 8ohm load.

120Hz hum, with input shorted: 1.5mV (with AC heating!)

1W freq range -1dB: 10Hz - 27kHz
1W freq range -3dB: 8Hz - 38kHz
2.5W freq range -1dB: 12Hz - 25kHz
2.5W freq range -3dB: 10Hz - 30kHz

The numbers actually go a little lower than I was expecting, but below 20 Hz the sine wave starts to look distorted. Can't be too sure of the low end numbers since my little Tek scope and function generator aren't at their best down there.

Overall I'm really pleased with the outcome. I could go on and on about the large sound stage or how vocals emerge from the blackness, isolated in space... but I'll spare you. And those V-caps aren't even fully broken in yet!

I do think a big part of the excellent sound comes from the use of high quality iron (and nickel too). Thank you Mike!

Dan

MagneQuest/Peerless Forum

MagneQuest/Peerless Forum

latest amp build

Post a Message!