1. Things you need to know about the new ‘Conversations’ PM system:

    a) DO NOT REPLY TO THE NOTIFICATION EMAIL! I get them, not the intended recipient. I get a lot of them and I do not want them! It is just a notification, log into the site and reply from there.

    b) To delete old conversations use the ‘Leave conversation’ option. This is just delete by another name.
    Dismiss Notice

Superreg on a Hackernap

Discussion in 'd.i.y.' started by john.luckins, Sep 11, 2013.

  1. john.luckins

    john.luckins pfm Member

    Two weeks ago Stefan (orangeart) came over to my place and we had a listen for a second time to his Hypex UCD400 amp and SMPS supply on my Buffalo system. Once again it put my well developed Hackernap a little in the shade and left me wondering what could still be done to further improve my Hackernap. Stefan was kind enough to leave the Hypex with me as he was unable to use it for a while. In doing this he inadvertently increased the pressure on me to do something for the Hackernap. I'm now very pleased you did leave it with me Stefan.. thank you.

    At that point my Hackernap was already quite well developed. Specs were as follows:

    1200VA Canterbury Windings output stage transformer with separate windings for +ve and -ve rails.

    A single hackercap for both channels.

    High voltage R Cores feeding choke regulated PSU with a further two LC filters before input to the FE supply to both Hackernap channels.

    Transformer on signal inputs to interface with balanced SEN supply from my Buffalo II.

    As I had been playing around with a multitude of regulators to power my Buffalo and SEN to great effect I decided it was time to add further regulation to the front end supply of the Hackernap. The current rating of the chokes meant that I couldn't use a shunt regulator such as the Salas. The Flea was/is a possibility but I already had a few spare ALWSR's and decided to try these.

    The easiest way to integrate an ALWSR into the feed without surgery to the Hackernap boards themselves was to fit it with the final FE choke as its load (largely resistive). I regulated the opamp supply on the ALWSR and fitted it to feed the positive rail of both channels of the Hackernap with approx. 53 volts.

    To be honest I didn't expect to hear a great deal of improvement but I was in for a surprise. The difference the ALWSR makes to the already fine Hackernap is not subtle and has no downsides that I can detect. My son noticed something was up and he wasn't even in the same room. Two weeks on and after some further tweaks to the ALWSR and the addition of a further ALWSR on the negative rail as well as additional gyrators before the ALWSR, I'm listening to a totally different animal of an amp.

    Yesterday, for a second time, I compared it to Stefan's Hypex. The Hypex sounds good of course but the super-regged Hackernap is just more enjoyable to listen to.

    In due course I'll add full details of the circuit and modifications to the ALWSR's for this application here on this thread. I plan to replace the gyrator on the Hackernap with the ALWSR next as I believe it would be best to have it right up against the front end rails of the amp, and have gyrator(s) further upstream in the power supply.

    Perhaps it would be a good idea for someone else to try this on their Hackernap. They would need to be able to release at least 6 volts of headroom which could come from reduced resistance in the RC filters on the front end supply, and a stereo hackernap with just one FE supply would be easier to try this on as it would only need one modified ALWSR.

    Anyone fancy trying it?

    John
     
  2. timH

    timH Senior Moment Member

    Well I would but I'm building four HackerNaps for active SBLs and I went and sold off all my spare ALWSR boards a few months ago and as far as I know they're not available anymore
    Tim
     
  3. mr_phil

    mr_phil If it isn't broken, try harder

    John, I am always up for a bit of DIY and I think Sam is bringing / lending his hypex this weekend to give a comparison
     
  4. orangeart

    orangeart KJF Audio Ltd.

    I can't wait to hear it John, I haven't got any further with my mega dsp dac yet so it will be a while. Did you get the hypex into balanced mode?
     
  5. Bemused

    Bemused What's this all about then?

    Top class work John.
     
  6. sam_cat

    sam_cat ᶜ ᶦˢ ᶠᵒʳ ᶜᵒᵒᵏᶦᵉ

    Lending, all ready to go :)
     
  7. Pete MB&D

    Pete MB&D Pete Maddex, the one and only!

    Mmm I have some teddyregs kicking about some where and 2 STRs, but need 4.

    Pete
     
  8. Bemused

    Bemused What's this all about then?

    Its a job for a Flea :D
     
  9. sam_cat

    sam_cat ᶜ ᶦˢ ᶠᵒʳ ᶜᵒᵒᵏᶦᵉ

    Got some flea boards here... Will bring them as well.
     
  10. 337alant

    337alant Negatively Biased

    Excellent write up John

    Tony can a flee take the high voltage?.

    Alan
     
  11. Bemused

    Bemused What's this all about then?

    Not a standard flea Alan, needs some bodging around also converting a flea for the negative rail is a bit of a phaf to be honest and involves carefully drilling out a via.
    In hindsight a flea can do it but not straight forward.
    At least the ALW has a negative rail board.
     
  12. 337alant

    337alant Negatively Biased

  13. Pete MB&D

    Pete MB&D Pete Maddex, the one and only!

  14. flashgo

    flashgo DIY Practitioner

    Wow, looking forward to visuals. I've got spare ALWSRs and other regs, so this could be just the kick I need to finish a HackerNAP. Also, I wasn't looking forward to the cost of the Hypex modules. Thanks John.
     
  15. df_genius

    df_genius Solder slinger

    I've already started on a similar path. I got my 2nd HackerNAP to around 90% in a fairly standard build but decided to go back and re-do the front-end regulation. Until reading this thread my plan was TX -> CRCRC -> Gyrator -> Acoustica'd TPR. The transformer here is a potted 400VA with two sets of 45-35-0-35-45 windings from a Harman Kardon Citation power amp

    I also have HackerNAP no.3 waiting in the wings, with two 250VA 45-35-0-35-45 Holden and Fischer toroids (probably from 1970's Crimson kits) that I thought would be ideal for HackerNAPs and was planning on using Coffin op-amp regs (similar to Jung) using OPA551 which are high-voltage op-amps. Using these toroids I have around 13-14V headroom for the front-end, and was also thinking of using a Gyrator before the coffin regs on these.
     
  16. john.luckins

    john.luckins pfm Member

    That was quite a positive response. I had been holding off from posting about the benefits of adding the ALWSR because I have so little time to draw up the schematics and give full details/picies. I'll have a go at all of this this weekend.

    To deal with your suggestions/ideas:

    While I'm sure the Teddyregs will improve the front end as well I don't expect them to be quite as effective as the feedback based regs. The Teddyregs predominantly use RC filtering and have less line regulation at very low frequencies than the opamp based feedback regs like the Jung, Sulzer and Flea. The get around 60 db of LF rejection whereas the ALWSR gets towards 100db of LF rejection.

    To operate at these high voltages, the Flea would need to have a floating supply up and running before the incoming voltage rises above 36 volts. It would also need a higher drop out voltage to work properly. That said the Fleas full use of the AD797's output linearisation at high current outputs is tempting. One modified Flea per rail with a higher voltage input could possibly better the ALWSR in this application.

    The main benefit of adding the ALWSR was to the positive rail of the Hackernap. The signal related variations to the Front end current draw are almost only on the positive rail. Its front end negative rail current draw is only through constant current sinks and their current setting resistor. The effect of adding the ALWSR to the negative supply was much less dramatic than it was to the positive.

    Shunt regs aren't ideal if you are short of supply voltage and/or have series resistance in the front end supply as they have to sink a relatively high amount of current to operate well. By relatively I mean in comparison to the current drawn by the load (10's of mA) The Salas reg needs at least 60ma to be shunted to operate at its best. The ALWSR circuit uses between just 10 and 20ma for its housekeeping. The examples of shunt regs that have been posted would probably do a reasonable job but they will have poorer line regulation then the ALWSR and be more noisy because of the zener. High voltage opamps are a real option though they do not perform as well as the lower voltage AD825/AD797 or LME47910 that can be used in the ALWSR.

    Finally for now, I've tried quite a few things out to improve the ALWSR circuit. In particular I've focussed on filtering the feeds from its output to its voltage reference and the opamp supply as well as filtering the CCS supply. A teddyreg type of filter circuit proved the best sounding in these roles and it became clear that a clean LF was important to these.

    As much as I have time for at the moment I'm afraid but more hard facts and picies at the weekend along with some questions for potential guinea pigs!

    John
     
  17. df_genius

    df_genius Solder slinger

  18. john.luckins

    john.luckins pfm Member

    I've used the Tracking pre reg of an LM1086 (an LM317 equivalent) and used all the mods on the Acoustica site, in particular I've replaced the lower of the setting resistors with LED's and added series resistance to the cap on the output to ensure stability. I've also kept the current that sets the voltage up above 1ma. I added all these changes progressively and listened for changes as I went. I also increased the voltage across the tracking regulator (using extra LED's) to give enough voltage at its output to allow for additional filtering of the voltage feed to the CCS on the ALWSR.

    I have tried with and without the Tracking pre reg and in this application I prefer having it in place. I like the way it cascodes the output transistor of the regulator that follows it and have opted for an additional gyrator before the tracking regulator on the Hackercap and on my SEN supplies where there is enough spare voltage headroom. I'm aware the the tracking pre reg may introduce some noise and have tried, unsuccessfully so far, to replace it with a sort of tracking gyrator. There were start up issues.

    There appear to be audible benefits to a series of a stages of filtering and regulation as long as each is well implemented and terminated. This approach raises issues of the order and type of regulation and filtering as well how many stages are enough. Even after fully tweaking the +ve rail ALWSR and its tracking reg, there was still and audible improvement from preceding it with a gyrator similar to that on the front end of the Hackernap circuit. This improvements was much smaller than that of fitting the basic ALWSR in the first place though.

    [​IMG]

    Here's a working schematic from when I was just about to add the gyrator before the TPR to give a clearer picture. Where the asterisk is by the TPR is where I was just about to add the cap multiplier (gyrator) and at the bottom is the indication of the voltage drop I expected for this. In pencil at the top you can see the voltage output of the TPR was increased later on to 67.2 volts to give headroom for filtering for the ALWSR CCS. Notice how much more current is drawn by the negative rail (due to the current in the biasing string of R14 and R36 on the Hackernap) and the additional 21mA housekeeping current of the ALWSR. Notice also that he first two sets of 6800uf caps are 100 volt rated. This might just be enough to save them from overvoltage if I ever powered this all up without the Hackernaps connected.

    John
     
  19. GWM

    GWM Lost in the 70's

    Hi John Just seen this new thread I have some spare AWSR boards 2 +- each and have allready implemented some of your ideas the front 40H chokes on the front end supply to great effect.So I am in for trying this out,unfortunatly I had to go in to hospital for an emegency op and have only just got home so need a few weeks to recover but as soon as I can I will try it out.Looking forward to getting more details have to say love my hackernaps there may be beter out there but they just play music the way I like it.
    Geoff
     
  20. john.luckins

    john.luckins pfm Member

    Likewise Geoff, I would wager that you have some further improvements to look forward to with those ALWSR's and from what you say you have probably got the voltage headroom to play with as well. If you want to prepare a bit a sketch of you're front end supply now, rather like the one of mine above would be useful in figuring out how best to install at least the first ALWSR, probably the positive one first off. Here's to a hasty recovery!

    John
     

Share This Page

  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice