DIY talk

Discussion in 'DIY' started by Cspirou, Jul 25, 2021.

  1. Pancakes

    Pancakes Friend

    Pyrate Contributor
    Joined:
    Aug 13, 2020
    Likes Received:
    1,417
    Trophy Points:
    93
    Location:
    Atl
    Really??? Wtf man....
     
  2. Gazny

    Gazny MOT: ETA Audio

    Pyrate Contributor
    Joined:
    May 11, 2020
    Likes Received:
    2,209
    Trophy Points:
    93
    Location:
    open sky
    You could have transformers up the wazzo if you enjoy their properties. Endgame in terms of design varies from designer to designer. DC, OTL, Cap couples, passive, active, parts cost (in either direction), part count. some other considerations what amp you will be feeding into too.

    I have a TVC preamp here and it's sound isn't for me. but I could see someone else falling in love with it. Not enough bite for me personally.

    The B1 nutube seems pretty exotic, but a friend of mine loved the non nutube-b1, maybe even the pmillet buffer too.
    [​IMG]
    [​IMG]
     
  3. Ken Tajalli

    Ken Tajalli New

    Joined:
    Jul 28, 2022
    Likes Received:
    11
    Trophy Points:
    3
    Location:
    London, UK
    I have posted this elsewhere on the internet, so I thought, why not here.
    I am a tinkerer! so DIY goes with that.
    I had an issue, because there was no commercial solution available, I decided to do it myself.
    I have a Chord Hugo2 for my portable use, I also had a Mojo too, both use micro USB connection.
    The requirement:
    - about one meter.
    - not too thick to take too much space in the little case I have.
    - not too thin, so it would last.
    - soft and flexible.
    - low impedance, so if I use it to charge my Hugo2, it would go into fast charge mode (2A capable)
    - to have a ferrite core, and be shielded and low noise, but have the right bandwidth for Hi-res and DSD, otherwise, the Hugo2 may go into white-noise of death mode. (If you have one, you know what I am talking about).
    - have a right angle micro USB (so I wouldn't fiddle which way is up) and USB-C OTG on the other side to go into my phone, without hassle.
    Added problem:
    - I have a Huawei mobile phone. It has 128GB of internal storage, and room for additional storage card, but Huawei uses its own type, not microSD card. The cards are limited to 256GB max. That is not enough for what I wanted to carry with me, some of it DSD and Hi-res.
    I wanted more storage, but neat and tidy and small.
    The solution:
    My DIY cable !

    [​IMG]

    [​IMG]

    - The cable is about a meter long,
    - 3mm thick.
    - soft and flexible.
    - 24/768 and DSD 256 play flawlessly.
    - Hugo2 does indeed go into fast charge mode, so it is low impedance.
    - The USB plugs are as tiny as they can be, but strong enough to lift the phone or even Hugo2 off the ground gently.
    But that is not all, look closely at second picture :
    - Ferrite core is 7mm thick, and it has common-mode rejection winding on the ground and +5V lines, should help combat RF noise pick up.
    - The small box! is in fact a USB-2 Hub and 500GB of flash drive, inside a tiny copper plate enclosure, that acts as noise shield and heatsink. Once plugged into my phone (any Android phone or tablet or computer) both the storage and the DAC are recognized. The hub is fast enough to allow 35MB/s read and write, but the entire contraption, under normal use, burns only 80mA of power. This is important, since it is the mobile phone that is supplying the juice.

    Has worked flawlessly for the past 6 months, can coil it up to a tiny bundle when not in use.
    With a tiny adaptor for the USB-C can connect it to standard USB sockets, or any charger.

    I just thought I share.
     
    Last edited: Aug 28, 2022
  4. peef

    peef Friend

    Pyrate
    Joined:
    Oct 5, 2015
    Likes Received:
    311
    Trophy Points:
    63
    Location:
    Montreal
    Is "unnecessary" not code for "you should totally do it!" in audio? :) Transformer coupled pres were not uncommon in the early days.

    The Raven preamp is a design worth knowing about. It has both input and output transformers. I'd suggest you could do a bit better by using either a voltage reference or a current source on the shared cathode, thus eliminating the ultrapath cap.

    http://www.nutshellhifi.com/triode1.html

    Circlotrons are really complicated. They need a balanced input signal, floating supplies, and some means of keeping DC off of the outputs. Historically, it would use an output transformer for this purpose.

    With that in mind, it's worth asking oneself what problem this circuit solves.
     
    • Like Like x 1
    • Epic Epic x 1
    • List
  5. peef

    peef Friend

    Pyrate
    Joined:
    Oct 5, 2015
    Likes Received:
    311
    Trophy Points:
    63
    Location:
    Montreal
    And they're stuffed. I uncovered a lot of issues in the process, so will be revising the boards once more. I'm really pleased with performance so far despite this, but had so many unusual failures along the way that I thought this would be worth sharing.

    First off, the good. The input buffer and Vbe multiplier work great. The 2SK3557s bias the VAS right about where it needs to be, making it a good choice for the input, and doesn't require matching. The servo also works remarkably well, when correctly populated. ...That's about it.

    I screwed up the footprint on one of the matched dual transistors, and so the VAS biased to about 100x the current I would have liked. when I powered it on. For prototyping, I figured why not try some unmatched through hole parts?

    [​IMG]

    (What could go wrong?)

    These seem to do a bit better than the matched NSS parts (perhaps higher hfe and Early voltage, lower capacitance?), and the matched parts are getting scarce anyway-- so moving forward, TH it is.

    The VAS would also oscillate at about 9MHz (a new personal record, last time was just 3MHz!). I tried installing ferrites on the cascode transistor bases, which yielded a full power 17Hz squarewave output right into the load. I'm not sure what could have caused this.

    [​IMG]

    Switching to resistor base stoppers solves the problem. I've since tested it with values spanning 8R to 250R and haven't seen any oscillation, though the response does rise a bit at 3MHz. I'll be digging a bit more here as this suggests stability might still be marginal.

    After all the futzing, the current out version works quite well.

    [​IMG]
     
    • Like Like x 3
    • Epic Epic x 2
    • List
  6. dBel84

    dBel84 Friend

    Pyrate
    Joined:
    Sep 27, 2015
    Likes Received:
    3,229
    Trophy Points:
    113
    That is quite a creative fix :). It is just as well work got a little hectic and I haven't had a chance to dive into this yet. I am assuming that there is no way to fix the smd footprint and new boards are in order.. dB
     
  7. peef

    peef Friend

    Pyrate
    Joined:
    Oct 5, 2015
    Likes Received:
    311
    Trophy Points:
    63
    Location:
    Montreal
    Afraid so. The good news is that the part with the incorrect footprint is part of the servo loop; it does not need to be populated for the current feedback (voltage amp) implementation, only for the transconductance amp. Provided you install the base stoppers, it should work. But let me test it first. :)
     
  8. Cspirou

    Cspirou They call me Sparky

    Pyrate
    Joined:
    Sep 27, 2015
    Likes Received:
    8,200
    Trophy Points:
    113
    Location:
    Northwest France
  9. Pancakes

    Pancakes Friend

    Pyrate Contributor
    Joined:
    Aug 13, 2020
    Likes Received:
    1,417
    Trophy Points:
    93
    Location:
    Atl
    I use that filter in Pi2AES units and a couple of amps. It's rated at 36V and 3 amps if I remember correctly? So beefy enough for anything desktop.
     
  10. Cspirou

    Cspirou They call me Sparky

    Pyrate
    Joined:
    Sep 27, 2015
    Likes Received:
    8,200
    Trophy Points:
    113
    Location:
    Northwest France
    I meant more in hardware than power handling. The Noise Nuke has a 6 milliHenry inductor. The diyaudio filter combines to 4.4 microHenry (or 0.0044 milliHenry) total inductance.
     
  11. peef

    peef Friend

    Pyrate
    Joined:
    Oct 5, 2015
    Likes Received:
    311
    Trophy Points:
    63
    Location:
    Montreal
    I contend it's not the size of the inductor that counts, but how you use it. :) Here's my version of the circuit. I would make some changes if building it today.

    Mark Johnson also posted a SMPS filter for his Noir amp, which includes common mode filtering. It's too bad this wasn't included in the P089ZB board. CUI has a great article on why CM filtering is a thing worth pursuing.
     
  12. randytsuch

    randytsuch Friend

    Pyrate
    Joined:
    Oct 13, 2017
    Likes Received:
    259
    Trophy Points:
    65
    Location:
    Los Angeles, Ca
    Speaking of Mark Johnson and "power stuff", I finally bought the parts for my quasimoto
    Lets you determine the values for a transformer snubber circuit with a scope instead of having to calculate.

    Going to try it out on the amp I'm currently building.

    Randy
     
  13. Pancakes

    Pancakes Friend

    Pyrate Contributor
    Joined:
    Aug 13, 2020
    Likes Received:
    1,417
    Trophy Points:
    93
    Location:
    Atl
    The quasimodo is a cool gadget. Unfortunately mine is still just a board and some loose parts lol.
     
  14. randytsuch

    randytsuch Friend

    Pyrate
    Joined:
    Oct 13, 2017
    Likes Received:
    259
    Trophy Points:
    65
    Location:
    Los Angeles, Ca
    I have the surface mount quasimodo, only thing available when I bought it. Assembled and used this weekend. Once I figured out how to adjust my scope, it worked great.

    Rs was lower than I expected at 18 ohms, but I looked at the results thread and there are others that get low R values.

    But I forgot to order the snubber caps, hopefully I can scrounge in my cap stash.

    I was actually planning to build two and sell the 2nd one, but screwed up my order and don't have enough caps to finish the 2nd one, so that will wait for another time.

    Randy
     
  15. Cspirou

    Cspirou They call me Sparky

    Pyrate
    Joined:
    Sep 27, 2015
    Likes Received:
    8,200
    Trophy Points:
    113
    Location:
    Northwest France
  16. randytsuch

    randytsuch Friend

    Pyrate
    Joined:
    Oct 13, 2017
    Likes Received:
    259
    Trophy Points:
    65
    Location:
    Los Angeles, Ca
    To clean flux after assembly, I take an "acid brush", and cut the bristles down to about 3/8 to 1/2" long.
    I did this outside yesterday, over a trash can.
    Pour some isopropyl alcohol on the board, then use brush to clean flux off. While still wet, dab with a paper towel to soak up dirty alcohol. Repeat as required.

    Its harder on the component side, but you can still clean some stuff off.

    I saw somewhere that you could use a toothbrush to clean, I tried it and toothbrush was too soft IMHO. The cut acid brush works well.
     
  17. peef

    peef Friend

    Pyrate
    Joined:
    Oct 5, 2015
    Likes Received:
    311
    Trophy Points:
    63
    Location:
    Montreal
    I recently got access to an APx515. It is uncalibrated, and I'm still learning how to use it, so take the following with a grain of salt. :)

    I've been using it to optimize the gm amp stage by stage. With some minimal tweaking, here's what's coming out of the VAS (a single ended bjt with some emitter generation, but no global nfb). Gold trace is the output, purple trace is the input.

    upload_2022-10-16_9-59-26.png
    upload_2022-10-16_9-58-29.png

    Test frequency is 420Hz. The 1khz and harmonics are from the SMPS that seems to make its way to the input, as seen on the purple trace. It's predominatly h2 until you push it, at which point h3 and h4 creep in at a much lower level.

    Solid state has a reputation for having predominantly odd order harmonics, so it's interesting to see how it performs in a single ended setup without high gain feedback to smooth things over. Next step is bringing h2 down a bit.
     
  18. Cspirou

    Cspirou They call me Sparky

    Pyrate
    Joined:
    Sep 27, 2015
    Likes Received:
    8,200
    Trophy Points:
    113
    Location:
    Northwest France
    nice
     
  19. Justin S

    Justin S Friend

    Pyrate Contributor
    Joined:
    Mar 14, 2018
    Likes Received:
    1,506
    Trophy Points:
    93
    Location:
    Toronto
    I'm a DIY noob. I make a lot of my own cables, and do my best to follow the occasional schematic to build something fun. I can solder OK: I am not going to burn things out but I'm certainly not going to win awards.

    I am excited to try out a crossover kit from GR Research on a pair middling Wharfedale 9.1s. I saw the kit, watched the videos and set out to look for a pair of these.

    I waited almost a year to find them - it's a sweet slow-burn project.

    We have listened to them for 6 months in our dining room and they're more than adequate for background music - which really is the use-case. I did have the P3ESRs in the room at one point. They were totally silly overkill there but brought a smile to my face every time I put them on. These Wharfedales...less so. They're good for what they are...but they are what they are.

    I am curious about how a DIY crossover with reportedly decent components, internal wire and a new design will change these. It's a low risk, coarse experiment but I have the time this weekend to finally give it a try. I want to see if I can bring these up a notch or two with a relatively small bit of $ and a few hours of my time.

    Tonight: disassembly and a well-themed, budget Saint-Émilion.

    PXL_20221112_022846127_2.jpg
     
    Last edited: Nov 11, 2022
  20. Justin S

    Justin S Friend

    Pyrate Contributor
    Joined:
    Mar 14, 2018
    Likes Received:
    1,506
    Trophy Points:
    93
    Location:
    Toronto
    Here is where I ended up with these. It took longer than expected for a few reasons:

    - Danny's crossover is massive and I had to split it onto two boards which needed to be interconnected while installed in the cabinet - this meant some fussy soldering inside the cabinet.
    - had to customize the rear panel for new binding posts
    - I second guessed my read of the schematic a bunch of times and had to keep retracing my steps. I am continuing to learn about translating schematics into components that are wired in parallel and components that are serial.

    Here's a few photos of the build. Yes - I scorched the backing with a torch when I was burning off the insulation from the leads on the inductors. I'm no pro.

    The stock speakers were pleasant, soft, a bit dull, and really served the purpose of "music in the room" just fine.

    My initial impressions with the new crossover:

    - more clear, more separation between the elements of the music.
    - transients are much more "crisp"
    - the speakers now disappear even though they are right against the front wall in an actual bookshelf.
    - They're a bit bright at the moment. This is probably partly new caps and partly my brain.
    - This system is a simple Chromecast Audio -> Toslink -> NAD D3020 MK2. With the improvement in technicalities in the speakers, I feel like I am hearing a bit more of the modesty of this system now.

    We'll see how these develop in the coming weeks. While it was fun to do, it was not a trivial effort for me. I'll pop in to update after I have lived with it for a while.

    One question I do have is about the dampening material that was originally stuffing the cabinet. I added the GR-Research "no-rez" foam to all of the internal surfaces. I did not replace the loose Wharedale dampening material back into the box. Should I?

    PXL_20221113_003906319.jpg PXL_20221113_010311539.jpg PXL_20221113_232455169.jpg PXL_20221114_223655193.jpg
     
    • Like Like x 1
    • Epic Epic x 1
    • List

Share This Page