Tuesday, December 15, 2020

nos TDA1541 tube DAC


I've become cynical of the latest is best mentality because in my 30 years of searching for sound that pleases my ears, I've always gone back to move forward. From this experience, I've learned to use planned obsolescence to my advantage by investing in old school components that have achieved classic status.

 There is a common denominator between the Philips TDA1541 DAC chip and the Western Electric 300B directly heated output triode vacuum tube. The former brought digital sound to everyone’s homes in the '80s, while the latter amplified talking pictures in movie theaters in the '30s. Both have also been rediscovered for their distinguished musical abilities.

This tubed TDA1541 sans oversampling multi-bit DAC evolved from my frustration with its bitstream and delta-sigma successors, revisiting and studying multi-bit CD players and DACs, performing Magnavox CDP + DAC mods, and other digital audio experiments.

!!!WARNING!!! 

The voltages in this circuit are potentially lethal! Build at your own risk!

If you are a metrics-obsessed audiophile who spends more time arguing in audio forums instead of listening to music because you can't trust your own ears and would rather buy a component approved by a hifi guru and/or his/her test equipment, STOP 🚫! You'll be offended by the contents of this blog, especially this project!

However, if you agree with Radiotron Designer's Handbook, 3rd edition published in 1940, that the THD metric should be abolished and better test procedures and standards established, you may proceed.😊

This project starts with the deconstruction of an already obsolete DAC board still available from eBay sellers. Since the TDA1541 chip has been out of production for over 20 years, I scavenge them from the CD player graveyard.


I incapacitated the analog section of the DAC board by removing the high speed/fast slew rate op-amps and low ESR signal capacitors, which will be replaced by a slower and harmonically more satisfying circuit from a bygone era. 

Non Over Sampling (NOS) measures like a broken component but my ears prefer musicality over accuracy, so the SA7220P chip was tossed and the circuitry bypassed. Speaking of accuracy, to what? I wasn't at the recording session, so I don't obsess about it.πŸ˜‰


This picture shows the spots where I tapped the right (AOR 6) and left (AOL 25) analog outputs of the TDA1541 chip. The DAC's delicate and jagged left and right analog output stream will be connected via shielded cable to a cutting edge I/V converter + vacuum tube smoothing circuit advertised and marketed by savvy Mad Men as retro-tech art.😎


I was very impressed by the sonic flavor introduced by the Western Electric 396A/2C51 in the buffer section of my 21st century Xiang Sheng DAC 01A. But in this particular application - gain stage + buffer -  the WE396A couldn't deliver the Carnegie Hall and Village Vanguard atmosphere I was trying to recreate. Caveat: the yellow WE logo doesn't offer a 100% satisfaction guarantee. 


So I rewired the B9A 9-pin sockets for a pair of 6CG7s, which I've already used successfully as a buffer for my TDA1543 DAC. These are the Noval-based equivalents of my favorite medium gain (mu)/medium transconductance (gm) 6SN7 octals, esteemed for their linearity as an audio amplifier. 

The circuit topology is very similar to my line stage preamp with the first triode section configured as a gain stage direct coupled to the second triode section wired as a low impedance cathode follower/buffer. No negative feedback is applied in the circuit to maintain signal purity. This tube circuit also functions as a sound processor, which is beyond the abilities of fast slew rate op-amps.πŸ˜„


I recycled a power transformer from a Heathkit mono FM tuner. Using a 6X4 tube rectifier, the rectified 480VCT  @ 60mA secondary feeds a capacitor input filter, then the B+ rail is split via 2 x 15H @ 30mA chokes to provide 295 volts of ripple free DC to each channel. The 6 volt DC supply, which lights up the 6CG7 filaments, is provided by a separate filament transformer. This archaic method of building a power supply enhances the DAC's tonal character!


Finally, I assembled the Back to the Future DAC on an organic 3/4" thick Baltic-Birch bread board. Even if it has a USB input, DSD, MQA and other so-called Hi-Res files are downsampled to 16-bit/44.1kHz PCM. That makes it 100% compatible with at least 90% of music streamed on the web. Since it makes my two-for-a-buck 4.75” silver discs from thrift stores and flea markets come to life, I don't ask for more.

Let's have some chips...



The reputation of this ubiquitous DAC chip was built upon regular production run TDA1541s, which can still be acquired for under $50. BTW, this is a DEM (Dynamic Element Matching - see TDA1541 FAQs) chip, not an R2R ladder DAC chip, as it's commonly misidentified. 

So there's no reason for angst if hand-selected and tested single and double crown S1 and S2 TDA1541A chips command SothΔ“By's prices or reside in private audio museums in Asia. Buyers should actually beware of high dollar re-stenciled regular TDA1541s studded with crowns. They will work but won't sound any better than the chip I meticulously unsoldered from an '80s junked Magnavox CDP.

1kHz square wave from Stereophile Test CD3
Stock Magnavox CDB560 CD player

Fortunately, I got a chance to sample a genuine single crown S1 graciously donated by Gary B. This select chip made my DAC sound a bit more transparent and cleaner, with a touch more detail and resolution. But I have to reaffirm that not having an S1 isn't a deal breaker. In fact, the darker and warmer sounding regular TDA1541s and TDA1541As can be a blessing in systems that are a bit hot in the top end. 

1 kHz square wave from Stereophile Test CD3
nosTDA1541tube DAC

 Not very relevant to this NOS DAC project, but AFAIK, the only difference between TDA1541A and the non-A is that the A can do 8x oversampling while the non-A is limited to 4x. But let's not go there anymore. Been there, done that...


NOS and Digital Filters


Before I even became aware of NOS DACs, I had been using line transformers as digital sound processors. If you're worried about non-oversampled digital artifacts destroying the irreplaceable diaphragms of your Altec 3000 super tweeters, look for bandwidth limited military surplus transformers. 



But the true analog-sounding line transformers JE Labs uses are those that have a flat response from 30 to 15,000 cycles, like the bokeh'ed pair of Western Electric 111C repeat coils and the UTC A-20s in front. There are so many sleepers out there branded with more letters and numbers, so don't limit yourself to UTC or WE.

My Analog Distortion Generating DAC



I’m happy to report that this repurposed '80s state-of-the-art digital music maker using Mid-Century Modern era devices and sensibilities has been fulfilling my analog-biased musical expectations at the end of this second decade of the 21st century. The DAC bread board doesn't exactly match the LeCorb and Mies aesthetics of my man cave, but I've been enjoying digital encoded music like never before, such that I can't get myself to rebuild it on a nicer chassis.

Many thanks to Gary Bronner and Pete Boser for their contributions to the JE Labs digital audio museum, as well as to the patrons of this blog!


πŸŽ„Happy listening and have a great holiday season!πŸŽ„






Tuesday, December 1, 2020

Home Brewer: Silbatone 755 Cabinet DIYers Club

Left = JE Labs clone (by John Piro)
Right = J-Rob's Silbatone original 

"The cabinets were designed by genius Silbatone engineer Dr. Stefano Bae as a group project for the Silbatone-sponsored DIY/Western Electric forum in South Korea. 100 pairs were quickly sold out, immediately driving up the world price for 755As. The original uses Silbatone's custom silver litz wire, which is said to be an essential ingredient of the original recipe, and top shelf WBT connectors.

Over the years Silbatone designed and made literally dozens of experimental boxes, including back horns for up to four (755A) drivers per side. This was probably the most successful, although rather modest.

My analysis of this untuned little cabinet is that it falls under the 'small box with a hole in it' design paradigm."

May 17, 2019



After I uploaded the Silbatone cabinet DIY info, I've received numerous inquiries about what drivers to use on this cabinet. Since it's impossible for me to try all the 8" drivers ever made, I thought it was best to let my readers chime in with their creations.

Altec 755As in Manila



Tango MCT999 MC transformer,
JE Labs 5691 passive EQ RIAA phono + 76 > Tango NP216N line stage, 
JE Labs SE2A3 with Magnequest TFA204 OPTs (on the floor)

Nonong is the only high school classmate and friend I remained in contact with after leaving Manila many moons ago. He coined SETUP (Single Ended Triode Users of the Philippines) back in the '90s and was one of the three pioneers listening to idler driven turntables, high mass tonearms, SPU cartridges + transformers, DIY preamps and SET amps driving high efficiency speakers.

Garrard 301 + Ikeda IT407 + SPU Meister Silver
JE Labs SE300B deluxe mono blocks with Tamura F7002s 
on copper chassis

His cherished pair of Altec 755As have been mounted on open baffles since the mid '90s. But as soon as he saw the Silbatone cabinet plan, he had the cabinets built in no time.

Altec 755As in the USA


Nate Grindheim aka Redboy in Audiokarma and HiFi Haven is an avid DIYer. You'll find his meticulously crafted MC step up transformers and amps in those aforementioned audio forums as well as in his Flickr stream.

Nate says:

"My buddy Gary (aka ejfud) built the cabinets for me. He knew I had the drivers and got tired of waiting around for me to build something for them, so he took the initiative to knock a pair together for me. He's a good friend to have around.

These things are really wonderful. They have that amazing Tone (with a capital "T"!) that I'm after, and the sound... it's just right, and good. They make music.

I am excited to get some proper amplification for them. There's a pair of 6L6 IPC amps here with Peerless iron that I think will be a great match, but I still need to overhaul them."


Altec 755C in Manila



Ben C recently retired from banking so he can focus his energy on his DIY audio passion. He's been stuck in Toronto since the Covid 19 lockdown and is itching to return home to finish his other audio projects.

Garrard 401 + RMG309 + FR64S + SPU Aniversary,
DIY preamp and SET amps

 I'm hoping that he'll drop by before he heads back to Manila when this is all over.

Altec 755E + JBL 075 in Manila



This bespoke cabinet is Joey's interpretation of the Silbatone executed in the Harana Audio Workshop


Harana Audio is also the official representative of Shinichi Suzuki and his Western Electric inspired G.I.P. Laboratory field coil drivers in the Philippines.

Lafayette SK98/Pioneer PIM8L in the USA



I've featured this driver and cabinet in a full-length profile of John Piro's creations. But it's worth reiterating a non Altec 755 driver successfully used in the Silbatone cabinet.


If you want your DIY Silbatone cabinet project added to this exclusive club πŸ˜‰, please leave a comment with your email address. I will reply to you directly.



Tuesday, November 17, 2020

John Linsley-Hood Class A Headphone Amp

Having been a happy user of a John Linsley-Hood (JLH) 1969 Class A amp cloned by Nobsound, I was intrigued by the JLH Class A headphone amp boards featuring discrete circuitry (no ICs or op-amps) advertised on eBay. In the midst of the 2020 lockdown, I ordered this $25 pre-assembled board, which arrived in a padded pouch within 10 days. I also got a surplus 24 VCT @ 2A power transformer for another $15.

!!!WARNING!!! 
The voltages in this project are potentially lethal! Proceed at your own risk!


Typical of Chinese kits, this board didn't come with instructions or any type of documentation. But the board is marked for an experienced DIYer to decipher and set up in no time. Unlike the original JLH 1969 design, which needed DC blocking capacitors at the output due to the single rail supply, this headphone amp module was updated to work with a dual rail (+/-) supply so the output capacitors can be omitted.
 

Unfortunately, I couldn't get stable zero DC offset at the output terminals. Trimming for zero volts after 30-45 minutes of warm up was not a guarantee that in another 30 minutes it'll still be zero - I've measured as much as 0.3V.😞


Since I didn't want to risk frying my headphones and classic speakers, I installed 2500uf @ 16V electrolytic blocking capacitors at each channel’s output. Then, I bypassed the electrolytic caps with hermetically sealed 4uf paper caps just like I did to the Nobsound NS02g.  

Top trace = audio generator
Bottom trace = amp output

In spite of the DC blocking caps at the output, the square waves still show excellent bandwidth, transient response with no trace of ringing. 

Before clipping, I measured 250mW/channel @ 4 ohms, 500mW/channel @ 8 ohms and 950mW/channel @ 16 ohms, 1.2W/channel @ 32 ohms,  395mW/channel @ 150 ohms, 190mW/channel @ 330 ohms and 115mW/channel @ 560ohms. I pretty much exhausted the resistor values from my parts bin to emulate the nominal impedance of typical headphones. 

The 115mW output into a 560 ohm load shouldn't worry headphone enthusiasts since this amp drove my Austrian AKG K240/600 ohm Monitor cleanly beyond SPLs my ears could tolerate. It also had no issues driving less efficient ortho-dynamic headphones like the Fostex T50RP MKII below, courtesy of my friend Nate. Thanks, buddy!


This amp definitely outclassed the Rat Shack mini amps I'm fond of. Its class A design's purity and transparency puts it on another league. Unfortunately, Class A amp designs require a more substantial power supply, which is incompatible with portability. 

The character of this amp is exactly like that of its bigger brother, a fatigue-free solid state amp that almost (but not quite) succeeds in sounding like a good tube amp!


BTW, did I mention that it can also drive my Altec 755As in Silbatone cabinets?