Your Current DIY Project Pictorial

The weird thing I encountered was that the silkscreening for my first set of boards was far superior to the second set (although those were probably more accurate to the design file).

 

This morning I swapped out the Jupiter copper foil caps in my ANK DAC 2.1 for some VCap Odams just for some fun.

Anyway, while I was inside the box I thought it worth taking some measurements because I've been considering a further upgrade attempt. I reckoned there may be visitors to this thread that may be interested in seeing what the output of this filterless DAC looks like.



These are a 1kHz signal measured at the input and the output of the tube analog stage. There's about 12mVrms of ugly waveform going in and a robust 2.7Vrms coming out. The tube gain stage looks to be inverting the signal phase also, which is something I hadn't considered. My tube preamp does so as well, so the chain ends up being in phase at the amp output (unless the preceding digital section outputs inverted too).

Compare the above sine waves to one from a NuPrime DAC-9.

 

Thanks for sharing this. For a scope novice, that's me, what is riding on top of the 1KHz input and output? Out of curiosity were the Vcap caps already burned in? Are you able to share your thoughts on the sound quality difference.

Another project I have been looking into is upgrading the guts of my Sony ES SACD/CD player. I think there are gains to be had there as well.

 

If you count them you could probably figure it out

Those are the audio samples at 44.1kHz from the D/A convertor, so there are 44 of them in the 1k waveform cycle. With no oversampling digital filter prior to the D/A stage, and no real reconstruction filter after, those are the artifacts left over at the 44.1k sample rate.

 

As @Davey says, it's what an analog output from a DAC without a digital reconstruction filter looks like. Basically, I think, each sample is held until it's updated with the next sample and that's what gives you the 'stair-step' effect. The reconstruction filter can do just that - completely reconstruct the waveform from the samples given.

It's amusing to note that filterless DACs of this nature have the reputation of sounding the most 'analog' when their output appears much closer to the traditionally incorrect view of digital audio as being 'jagged' and stepped like a staircase. Me, I like the sound of it.

Audio Note used to use a passive filter in their DACs, I believe, but gave them up claiming they hindered sound quality. I'd be interested in trying something like that though, if only to filter ultrasonic noise.

*oh, and I've not listened to the VCaps yet.

 

The I/V transformers do provide some filtering, and probably the gain stage too, but I don't know what the rolloff may be. There's obviously some shown on the scope, but not a lot.

 

I'm very curious to see what a different tube gain stage might do. I haven't gone too far into it yet, but on the face of it Broskie's Aikido circuit, with it's inherent noise rejection and perceived neutrality might be the perfect thing for a DAC output stage...

 

Yea, I'd like to try a transformer I/V convertor, could probably use one of the MC step up transformers I have around. In my DAC, I used the Burr Brown DF1704 digital filter in slow rolloff mode, and the current output from the PCM1704 convertor goes to a simple discrete I/V current mirror stage operating without feedback using heavily biased low noise bipolar transistors, providing about a 4 ohm dynamic impedance to the D/A convertor output, followed by a cascoded mosfet stage operating single-ended class A with no feedback and about 70mA bias. Sounds very good, simplified schematic below...

 

Not an overly complicated or exciting project, but possibly controversial…. I’ve recently started a cable company and have been building my silver cables for commercial sale. I’ve been wanting to be able to burn them in before I ship them so I built my V1 burn in rig. I’m using an old Yamaha M-85 amp for its ability to deliver significant current and have a 1000 watt 6 ohm resistor to apply a load. It’s a bit of a ruff and tumble build, but it’s entirely utilitarian.

 

I looked at the pictures before reading your message and thought surely that can't be a power resistor, can it?

How many hours does a set of interconnects spend on this?

 

The resistor is impressively large. I’m going to run them for 72 hours and then listen, but I’m not sure how much or little will be needed, or if it will even make a substantial improvement at all.

 

Apologies in advance for another post of oscilloscope pictures. I hope someone out there finds them interesting.

I have been messing around with low pass filtering for my ANK-style DAC and the results have been quite an eye-opener. My thought was that the easiest place to insert one would be an RC filter just before the grid leak resistor of the tube gain stage. It wasn't very successful initially and it took a little while to figure out that I had neglected to consider the resistance of the preceding stage, which I guess is the 1.2k winding resistance of the I/V transformers.

With that in mind the RC filter already had its R (resistance), so in theory all that was required was to increase the value of the 820pF capacitance on the tube stage input. I soldered a 0.1uF cap (the C) across the 1K load resistor in one of the channels so the two sides could be compared.



As before, the 1kHz tone in the unaltered channel displays the unfiltered stair-stepped waveform of the DAC output (left), but the channel with the added capacitor is smoothed flat. As a non-engineer, this is really amazing to me. It's almost like magic.

-----

Next I bumped up the frequency to 10kHz. The unfiltered sine wave is now quite ugly. It's surprising that the signal can be distorted to this degree and still remain listenable, if not, for some, preferable. Although the other channel remains smooth, the voltage has now dropped considerably due to the heavy filter installed. With such a large capacitor the frequency rolloff started before 1kHz.



-----

The 0.1uF cap was swapped out with a 0.01uF and the incoming signal frequency upped to 20kHz. More extreme sine wave distortion was now evident in the original configuration and some starts to appear in the peaks of the new filter also.

Going right up to 40kHz distresses both sine waves considerably (bottom left & right of the below image), with the filtered side now only about 1 volt (peak), down from 3 volts and forming a triangle wave. I believe at this point both waveforms may have been modulating.



If I've worked it out correctly the 0.01uF channel cutoff frequency (the point where it's 3dB down) is only 15kHz, but I will try and give it a listen and see how it is!

-----

Below I've done a quick & dirty chart of the frequency rolloff with three different cap values. The horizontal line is the 70.7% mark which I think is the -3dB point.



This has been a nice little learning experience, but I'm not certain it's a particularly helpful mod. A more complex filter configuration is probably required, if that'd what someone wanted. From what I've been able uncover Audio Note used a CLC filter (when they had one) on the DAC board itself and it doesn't appear the I/V transformers were utilised until the filter was removed.

 

Probably late to chime in, but you'd be wanting feedback components to be as close as possible to the inverting-in and out of the op-amp.

Here's one of the few IC pin layouts that makes it quite easy (the pins of a DIP8 are 2.54mm apart; this opamp is 3mm x 3mm), but the standard 8-pin x2 package is a close second.

 

You could try Ctotal values between 3900pF and 4300pF. This would give -3dB points of 41.5kHz and 37.7kHz (if my math is correct) which is well outside the audio bandwidth and below the 44.1kHz source of noise

 

Not too late at all, I'm still tweaking the board a bit, replacing the voltage divider with a TLE2426 (at the recommendation of a user @ another forum) and also swapping to some physically larger caps for input/output coupling to make routing with a single 0.8mm bit a little easier/nicer.

The diagram you posted doesn't seem to match the pinout for a typical dual op amp? May I ask where you got that?

 

OPA1622 would be one ... https://www.ti.com/lit/ds/symlink/opa1622.pdf

 

Thank you. Well I think @harby is right that I should tighten up my layout. Ironically, if my test boards were a little larger, I could snug things up considerably. My feeling was that I'd do the best I can using as much materials-on-hand as possible. If this doesn't totally suck, I'll get larger copper clad drops and move onto V2.

Edit to add: Nope, I do see how I can bring the RIAA loop in closer to the op amp in this current board size. So I'll do that. Thank you, @harby.

 

Thanks. Yes I think you're right, and with 3900pF it's only down 1dB at 20kHz.

 

Snugged things up the best I can, I don't think I can get the feedback parts any closer and still cut everything with a 0.8mm bit. I did look at some other boards with feedback-RIAA and my component spacing here now seems to be in the same ballpark. Thank you for your tip/help.

Here is a pic of the current layout, and below a pic of a previous test board in the blue enclosure I plan on using. I only ordered a single RCA jack when I was determining which jacks I'd be using (I'll need to order three more when I place my parts order).

The copper clad boards I'm using are just wide enough for the enclosure, but not as long. I had wanted to use the same sort of RCA connectors as Sutherland, which solder to the board PLUS there is a retaining nut to the enclosure, and that would hold my phono board to the rear panel and prevent it from sliding in when making the connections. BUT, four of those RCA connectors look like they'd cost 2-4x more than my entire project. Instead I plan on using two Keystone brackets to hold the phono board to the rear enclosure cover, the concentric green circles at the bottom of the board are the mounting positions for those brackets.

 

Looking good!

 

Breaking ground on a plinth build for my Technics SP-10. Starting with some marine grade Fir plywood, which is void free like Baltic Birch ply, but I’ll likely be making modifications as I listen to the table. The plan is to fill the bottom two layers with resin and lead shot, at least 50lbs worth. I’m going to mount an ET-2 tonearm to this that I will rewire with silver. Can’t decide on the connection for this yet, but I want to be able to run the cartridge as balanced or single ended.

 

Deeply, powerfully jealous.

 

You’re moving faster than me now! I’m trying to wheel and deal with someone who’s pretty savvy with a soldering iron to install the umbilical connectors for me so I don’t destroy an irreparable pcb

 

You’re talking about a new umbilical to get the electronics away from the platter I assume?

I got a wild hair and went into the shop a lot later than I typically would have. I managed to glue up a blank for the tonearm and motor. It’s a stack of 3 pieces and the middle piece already has the cut out for the motor. I’ll use a pattern trimming bit later to trim the outer 2 pieces of this sandwich to match the inner piece. I’m still up in the air as two how I’ll do an arm board, but that can be figured out later.

I also managed to put together the two pieces that will make up the bottom layers of the plinth. These each have 25lbs of lead shot , so 50lbs total, mixed with resin. After these cure out I’ll use the same resin to join the 2 together. I plan to mechanically fasten this portion to the portion that holds the motor and arm for now so I can easily have access to the machine screws that will be needed to hold the SP-10 drive unit to the plinth.

l’m expecting the entire plinth to check in at around 60lbs so with the SP-10 in the plinth the whole thing should be around 80lbs.

 

First two layers for the prototype/mockup finished and glued together to try fitting the Lenco chassis:







And finally after some adjustments I have started cutting the layers from birch plywood.




The pockets around the plinth will be filled with a mix of lead shots and an acoustic glue/compound like Tec-7. Which also will be used as glue between two of the plywood layers.

...and just a reminder what I am trying to build: