[rsabo]

I've drawn up some power supply boards for the Aikido headphone amp I've been planning/putting off for a year or so. I did this because a) I like the look of PCBs over point-to-point; and b) it'll help me avoid screwing up by, say, crossing wires, or mixing up the AC wires with the DC wires, etc.

I've done two boards; one power supply, and one regulator. The power supply is the one in the Aikido docs, the regulator can be seen here. (Go down to "A simple solid-state high-voltage regulator")

I'd appreciate it greatly if someone can point out any errors or improvements.

Power Supply:
Schematic:
http://img145.imageshack.us/img145/850/aikidops1km.png
Board:
http://img444.imageshack.us/img444/8...dopsbrd0tp.png


Regulator:
Schematic:
http://img215.imageshack.us/img215/3...kidoreg3sn.png
Board:
http://img215.imageshack.us/img215/6...oregbrd7we.png

 

[AndrewFischer]

Quote:

Originally Posted by rsabo /img/forum/go_quote.gif I've drawn up some power supply boards for the Aikid...

You've just done the next thing on my personal to do list. I'll give this a serious look later today.

At first glance, I'd make the traces wider.

The regulator schematic needs to be cleaned up, but that's cosmetic.

Using Eagle? Are you willing to share your .brd and .sch files?

 

[AndrewFischer]

Besides wider traces, I have a few more minor items

Rectifier board:

C5,C6 and C7 might be to close. Dunno. Depends on parts, but there isn't much room on the silk screen.

C5,C6,C7 and L1 are very close to the edge of the board. May be a non-issue. Depends on parts and case.

Have you thought about making the board wider and putting D5 between L1 and the capacitors? That might let you move that power trace.

Are the spaces on the ground plane wide enough? They look kinda small to me?


Cosmetic:

Rotate some of the resistors around so the silk screens don't overlap with other parts.


HV Regulator

There is an error in the schematic. The base of the TIP50 connects to C9 diode.

I didn't check the tubecad schematics for errors, so maybe you fixed something wrong in the original...

Your schematic is a mirror of the one in tubecad. If you draw it the other way, you won't have to cross the transistor base leads over and the schematic will be cleaner looking. It will also be easier to see errors. (minor cosmetic suggestion)

I'll take a look at the layout after you fix the error.


(if my comment are at all harsh or terse, I apologize. I've got a ton of work to do and wrote this as quickly as I could)


A.

 

[n_maher]

Glad to see some people discussing the Aikido here!

At the risk of exposing my inadequate knowledge what do both of you plan on doing for a heater power supply? It's the one thing on my Aikido that I have yet to really flesh out.

Nate

P.s. Here's my Aikido parts stash.

 

[AndrewFischer]

Nice stash. I plan to go cheaper on the passive parts.

Ever since I read Eric Barbours 1995 Glass Audio article, l've wanted to build an Octal headphone amp. The cost of building Barbour's amp with new parts kept me away. I've been considering building a mu-follower using NOS odd ball tubes. I've drawn up a few and done some SPICE simulations. NOS 8SN7s sell for about $4.00.

Then I read about the Aikido. Broskie's boards require both tubes to run the same filament voltage, so that more or less eliminates using odd balls like the 8SN7. Darn. I don't think his Octal boards can be configured as headphone amplifiers. Drat.

So I have to decide to either not use his boards or use 9-pin tubes. I'm 90% of the way to ordering a set of 9-pin boards from him.

Once I decide on tubes, I can figure out power supplies.

Assuming I go with 9-pin, 12V tubes:
12AU7 300mA 12BH7 900mA x2 for stereo or 2.4A @ 12.6V

Idea #1 is to use unregulated DC for the tube heaters. Use a STEPS board for the line filter, rectifiers and capacitors. Since theses are all indirectly heated tubes, the filaments are not in the signal path. Choose caps for their ability to kill 60Hz noise, not for overall frequency response. Blackgates are not needed.

I'd have to use an off-board transformer. That's more power than you can get from a 60mmX60mm Amveco transformer and too much current for an LM317T. I'd use the old trick of series connected diodes to cut down the DC voltage to close to what the tubes need.



Idea #2 is regulated filament power. Use two STEPS boards, one for each channel. 1.2A@1.6V is well within what I can get from a 60x60mm transformer. 1.2A is too much for an LM317 but not for an LM338.


At the moment I'm leaning towards #1. Traditional tube gear uses unregulated heater power and it works fine. DC should help with noise. As long as I leave room in the case, I can always swap to #2 if #1 doesn't work well.


There is also the possibility of using a step-down switcher. Maybe an LM3448????

 

[dsavitsk]

Quote:

Originally Posted by n_maher /img/forum/go_quote.gif At the risk of exposing my inadequate knowledge what do both of you plan on doing for a heater power supply?

If you only have a 6.3V secondary, you'll never get 6.3VDC out of it. Try AC before you do anything else. Just ground the center tap (or better, bias it up a few volts) and you should be fine. It is largely a myth that AC heaters on indirectly heated tubes are noisy. This is only the case if done improperly.

Quote:

Originally Posted by n_maher /img/forum/go_quote.gif P.s. Here's my Aikido parts stash.

For all the money you have spent on tubes (and the attenuator which will make almost 0 sonic difference), replace those solens with something decent. They really are garbage in the signal path.

 

[AndrewFischer]

Idea #3 -- Scale up the STEPS to fit the 35 and 50VA Amveco transformers.
I suspect a single LM338T will overheat with a PCB mounted heatsink. National has designs for using 2 or 3 LM338s to increase capacity. That should work.

I hadn't considered this as a new PCB wouldn't be cost effective for just me. But if might be for a run of 10 to 20 boards. For fun, I'll design a regulated linear suitable for the Aikido heaters. It will include a way to float the heaters to a fraction of the B+.


If 3 or 4 people are interested in splitting the cost, we can PM about getting a small run of boards made.


Edit: Might be impossible to fit on the 160x100mm maximum board I can make with the non profit version of Eagle. Any suggestions for a free CAD program?

 

[Nerull]

Just wondering dsavitsk, you said the attenuator will make almost 0 difference, I was wondering how you came to this conclusion? I'm curious, because I want to know if it's a big waste of money or not, because I guess I can't trust a lot of audiophile reviews, and I agree with your capacitor reviews (having tested capacitors of the same make myself).

Not to derail the thread or anything, I was just curious and have almost never seen a comment such as that.

~Tom

 

[AndrewFischer]

Quote:

Originally Posted by dsavitsk /img/forum/go_quote.gif If you only have a 6.3V secondary, you'll never get 6.3VDC out of it.

That's not quite true. Since 4 tube heaters are a more or less constant load, an unregulated DC supply will work.
With no load, the output will be about 1.414*6.3 - voltage drop across the diodes. Somewhere around 8VDC. As the load increases that will fall. A lot. How much it falls depends on the VA rating of the transformer, the regulation of the transformer and the size of the capacitors.

I do agree that AC is worth a try on indirectly heated tubes.

 

[dsavitsk]

Quote:

Originally Posted by Nerull /img/forum/go_quote.gif Just wondering dsavitsk, you said the attenuator will make almost 0 difference, I was wondering how you came to this conclusion?

It's a little bit of an over statement. I was just trying to say, in comparison to the capacitor choice, the amount of good it will do is small. Think of it this way, If I had some dollar amount to spend between coupling caps and a volume control, I'd likely put at least 75% of that into the caps. If you include power supply caps, up that to 90%. Yes, I know that the Aikido is supposed to be not very PS dependant, but I've heard plenty of other circuits that were not PS dependant be improved greatly by a better PS. So, I guess, for me, if the sonic benefits are the concern, and not the channel matching, or the nice clicky feel of the stepper, then I wouldn't be investing in a stepper until I had pretty high quality caps.

In Nate's case, I am assuming that he has super jumbo caps on the output to use it as a headphone amp. Even there, I'd likely use ASCs instead of the Solens, but whichever, I'd at least bypass them with something high quality -- Like a 1uF Auricap at a minimum. It will do more good than harm. But, I'd also try it both ways to see the distinction.

Anyhow, it is easy to test how big of a difference a stepper will make. Use a pot to set an amp at an appropriate volume, take it out and measure it, and use a pair of resistors to form a similar voltage divider. You can quickly, and cheaply, figure out whether you think it is an improvement. To my ear, I think a PEC pot sounds about as good as two resistors, and in fact sounds better than two metal film resistors.

-d

 

[n_maher]

Quote:

Originally Posted by dsavitsk /img/forum/go_quote.gif If you only have a 6.3V secondary, you'll never get 6.3VDC out of it.

I have a separate trafo for the heaters. Quote:

Try AC before you do anything else. Just ground the center tap (or better, bias it up a few volts) and you should be fine. It is largely a myth that AC heaters on indirectly heated tubes are noisy. This is only the case if done improperly.

Thank you for the advice, my concerns are mostly centered on how to ground the heater 1/4B+ as recommended and also how to adjust the voltage of the heater PS to make sure that the tubes are seeing something close to 6.3V.

Quote:

For all the money you have spent on tubes (and the attenuator which will make almost 0 sonic difference), replace those solens with something decent. They really are garbage in the signal path.

I respectfully disagree on all counts. One, the grand total of what you see there in tubes is less than $100, which doesn't seem like much for 13 tubes, does it? Two, I've experimented pretty extensively comparing pots to SA's and in every case I've thought the stepper sounded better, to my ears. That said I'm using the stepper because I already had it, it'd been sitting in a box for a year. Three, I've heard the Solens used as output coupling caps and thought they sounded nice. However, since I haven't actually started building anything I'd love to hear some alternate suggestions. I need something on the order of 47uF+, I couldn't find much out there that fit that description. I'd love to find some Sprague caps, but the Ebay supply hasn't yielded anything lately and no one stocks them.

Also, the high voltage power supply I'll be using is a custom design, tube regulated one so I agree 100% with you on that front. The three ASC's that you see in my stash are all for the power supply.

So if this seems like a threadjack, hopefully the Aikido specific discussion will be seen as helpful.

 

[dsavitsk]

Quote:

Originally Posted by n_maher /img/forum/go_quote.gif I have a separate trafo for the heaters.Thank you for the advice, my concerns are mostly centered on how to ground the heater 1/4B+ as recommended and also how to adjust the voltage of the heater PS to make sure that the tubes are seeing something close to 6.3V.

Just make a voltage divider from B+ to ground. So, use a 220K resistor, and a 71K5 resistor in series, and connect the center tap between them. It is also a good practice to parallel a small electrolytic (50uF or so, but what ever you have around will be fine, and quality does not matter at all, just be sure the V and heat ratings are high enough) with the resistor to ground (the 71K5 in this case.) Works like a charm.

Quote:

Two, I've experimented pretty extensively comparing pots to SA's and in every case I've thought the stepper sounded better, to my ears.

Fair enough

Quote:

Three, I've heard the Solens used as output coupling caps and thought they sounded nice. However, since I haven't actually started building anything I'd love to hear some alternate suggestions. I need something on the order of 47uF+,

I do like the ASCs marginally more than the solens, but since you already have the solens, I wouldn't throw them out. My experience is that bypassing the solens does a world of good. So, pick your favorite cap and put one that is from 1 to 2% the value of the larger cap (.47 to 1uF here) in parallel. I think PIOs work well here, but any high quality cap will work fine. Auricaps are pretty standard and are quite good.

Didn't mean to pick on your parts choices, Nate.

-d

 

[n_maher]

Quote:

Originally Posted by dsavitsk /img/forum/go_quote.gif Just make a voltage divider from B+ to ground. So, use a 220K resistor, and a 71K5 resistor in series, and connect the center tap between them. It is also a good practice to parallel a small electrolytic (50uF or so, but what ever you have around will be fine, and quality does not matter at all, just be sure the V and heat ratings are high enough) with the resistor to ground (the 71K5 in this case.) Works like a charm.

Thanks, I think I actually understand all that.

Quote:

Didn't mean to pick on your parts choices, Nate.

I sincerely appreciate you taking the time to have a look at them and will never take offense to someone else's point of view on things like this. It's not like there is a wealth of knowledge about the Aikido as a headamp to pull from (at least I couldn't find much) so it's great to hear some feedback about part selection.

 

[looser101]

Quote:

Originally Posted by n_maher /img/forum/go_quote.gif It's not like there is a wealth of knowledge about the Aikido as a headamp to pull from (at least I couldn't find much) so it's great to hear some feedback about part selection.

Latest impressions:
Hamilton mini-meet/ mod session with pics!!!

Good luck with your builds!

~Renato

 

[gsteinb88]

Quote:

Originally Posted by AndrewFischer /img/forum/go_quote.gif That's not quite true. Since 4 tube heaters are a more or less constant load, an unregulated DC supply will work. With no load, the output will be about 1.414*6.3 - voltage drop across the diodes. Somewhere around 8VDC. As the load increases that will fall. A lot. How much it falls depends on the VA rating of the transformer, the regulation of the transformer and the size of the capacitors. I do agree that AC is worth a try on indirectly heated tubes.

Unless im mistaken, i believe that math is wrong. To calculate the RMS voltage (that is, the DC voltage youre going to get off of an AC line) you divide the peak voltage (in this case, 6.3v) by the square root of 2, not multiply it. Youre never going to get a higher voltage than the peak voltage of the signal from a rectification circuit, even unregulated. So that would put the maximum DC off of a 6.3v center tap at about 4.47v.
Then again, i could be wrong, im in the middle of about 5 hours of studying right now (and being a bad student and nipping off to check head-fi

)