PANDA Headphone amp kit, build thread, upgrades & tweaks.

[joethearachnid]

Just a quick question:

At some point I'm probably going to be upgrading my main cans and I'm vaguely interested in getting a pair of orthos, since the HiFiMAN HE-4 can be had new for less than the HD 650 ($449).

However, as you probably all know, orthos need a lot of power. If I were to run a pre-out from the Panda through the back then through an adapter to the HE-4s (bypassing the output resistors) then by my calculations I'd be able to put ~3W through them (they're 38 ohms). Does this make any sense/would work?
I may be missing something with voltages/current here, but I was just wondering.

-JoetheArachnid

[Deleted]

Hi Joe,

Yes you could drive a 38Ohm headphone to just about 2W. (9Vrms) if you leave the output resistors out.

It won't be in full class A anymore but run mostly in AB.

[dicky]

Hi Mike, are you out there

I'm going to copy your Panda 4, out of interest and to use up a spare kit I've got. I'd be grateful if you could confirm some of the component values for me?

I've got all the electrolytics sorted and 22pF polyester replacements for the ceramic caps. Can you confirm the mica capacitor values (33pF?) for the diodes and the film (470nF?) and polystyrene (22pF?) decoupling capacitors?

Cheers.

[dicky]

OK, I used the values I mentioned in the previous post and she lives. I've set the voltages to 1.5V and trimmed to close to 0mV. I'll leave her to cook and check them again later. I'll report back once I've sorted some input/ouput cabling and had a listen - probably be next weekend now as I'm away in the week. It'll be interesting to see what difference the changes make.

[dicky]

Hi Solderdude, how can I tell what type of diodes are fitted? (I used the ones that came with the kit.)

I just measured 8.2V across one of the 1.5k resistors but I don't know what diode type I have fitted so I can't work out what value of resistor I should fit to optimise the zener circuit (1230 [1k2] or 676 [680]).

Could you help, please?

[Deleted]

You can tell by either the type number (if visible and then google this type number) or thickness.
There are also 2 smaller zener diodes these are 3.9V/0.5W I believe the slightly thicker versions are 1.3W.

Mick's Panda had 1.3W zeners in it so 680 will probably the right value.

[PinkFloyd]

Dicky,

I'm so sorry man, I have been working from 8am to almost 12 midnight the past while and haven't been that active on the forum. I think what you said above sounds about right.... 22pF polystyrene instead of the ceramics, the one on the diodes are fine and 470nF will be fine too.... have you fitted the 4700uF power caps?

Mike.

[dicky]

You can tell by either the type number (if visible and then google this type number) or thickness.
There are also 2 smaller zener diodes these are 3.9V/0.5W I believe the slightly thicker versions are 1.3W.

Mick's Panda had 1.3W zeners in it so 680 will probably the right value.

Hi Solderdude, thanks for getting back to me . My eyes aren't that great - I'm looking at the 4 (silver/grey) 18V diodes adjacent to the 1k5 resistors. All I can see on them is 4746. I googled '4746 zener diode' and the results returned '1N4746' and 1 Watt. So I guess that means the currently fitted resistors are definitely not 'optimal'. I can solder a 1k2 across each of the fitted 1k5 resistors to make 680 ohms rather than try to take them out.

Regards,

Dicky

[dicky]

Dicky,

I'm so sorry man, I have been working from 8am to almost 12 midnight the past while and haven't been that active on the forum. I think what you said above sounds about right.... 22pF polystyrene instead of the ceramics, the one on the diodes are fine and 470nF will be fine too.... have you fitted the 4700uF power caps?

Mike.

No problem, Mike. Yeah, I have 4700uF power caps. I connected 470nF and 22pF in parallel with them (copied your Panda 4). If I can confirm what zeners I have fitted I'll do Solderdudes 'optimisation' mod.

Dicky.

[Deleted]

I can solder a 1k2 across each of the fitted 1k5 resistors to make 680 ohms rather than try to take them out.

Yes, you can !
That would be the preffered option.

[PinkFloyd]

Hi Dicky.... 1K5 and 1K2 will give you 666 ohms (the number of the beast)..... here is a handy calculator for you: www.1728.com/resistrs.htm

[Deleted]

666.2/3

If you suffer from hexakosioihexekontahexaphobia another resistance value might be adviseable.
Should you consider this amp to be a 'beast' it is kinda suits ....

values between 600 and 700 Ohm will bias this zener correctly b.t.w. with about 8V across the resistor.
You can check this by measuring the voltage across the zenerdiode which should be quite close to 18V if biased correctly.

[gommer]

Is this zener bias advice already in the tips and tricks thread? If not, it should be.

Cheers, Marc

[Deleted]

this IS the Panda Build thread and sortof functions as a tips and tricks thread.
There is no separate tips and tricks thread (as far as I am aware of).

Every now and then I clean this thread up and keep the build related parts in here.

It is also mentioned in the bigger Panda thread b.t.w.

[PinkFloyd]

Feb 15, 2011 9:33:04 GMT @solderdude said:

666.2/3

666.67 according to another calculator...... 666.66 would have been better

[gommer]

Feb 15, 2011 12:29:44 GMT @solderdude said:

this IS the Panda Build thread and sortof functions as a tips and tricks thread.
There is no separate tips and tricks thread (as far as I am aware of).

Every now and then I clean this thread up and keep the build related parts in here.

It is also mentioned in the bigger Panda thread b.t.w.

Oh Oh, please clean up my useless remarks. I was confusing this thread with the "good value for money..." thread

[gommer]

Feb 15, 2011 12:46:18 GMT PinkFloyd said:

Feb 15, 2011 9:33:04 GMT @solderdude said:

666.2/3

666.67 according to another calculator...... 666.66 would have been better

Euhm, Mike, did you fail for maths in your younger days? 666 2/3 rounded to 2 digits is 66,67 and in fact it's equal to 666,66.. with an endless row of sixes.

I might also be missing the point if it was humour

[Deleted]

Feb 15, 2011 12:46:18 GMT PinkFloyd said:

Feb 15, 2011 9:33:04 GMT @solderdude said:

666.2/3

666.67 according to another calculator...... 666.66 would have been better

;D

[Deleted]

I will clean up this thread when I get the urge and all will be forgotten.

Am I seeing a wascal wittle wabbit with a purple AD700 ... on it's hat ?

my calculator says:
666.66666666666666666666666666667

meaning 666.6666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666666 e.t.c.
so 666 + 2/3 is the best fitting mathemetical description by lack of a 6 with a slash through it indicating the 6's go on endlessly.

I have my doubts about the tolerance of the resistors that this value could be reached in practice.. ;D

[dicky]

Hi Dicky.... 1K5 and 1K2 will give you 666 ohms (the number of the beast)..... here is a handy calculator for you: www.1728.com/resistrs.htm

Thanks Mike - that's a useful tool. I think I'll stick with 666 ohms

[dicky]

You can check this by measuring the voltage across the zenerdiode which should be quite close to 18V if biased correctly.

Solderdude, I just measured (from switch-on/cold) 7.6(ish) volts across a 1k5 resistors and 17.6(ish) volts across it's adjacent diode. All 4 resistor voltage values are slightly different. Perhaps they'll settle down when everything is up to temperature.

So by more than halving the resistor value 1k5 to 666, the diode voltage will only increase by ~0.4v?

[dicky]

Solderdude again! Sorry - perhaps I should shine a light in your face with all these questions! ................

another thing to tinker with is resistors R6/R17/R34R50 which can be deleted (shorted NOT removed) or made considerably smaller.
Now they are 1k but can be anything between 0 Ohm and 100 Ohm for instance.

NOTE: You should only mod this mod/experiment when you have properly adjusted the zener's bias current as explained above.

....... where do you start here? Just take a box of different value resistors and keep swapping them? Start at 0 Ohms and work up ....... start at 100 Ohms and work down? What kind of effects should be obvious?

Cheers

Dicky

[Deleted]

So by more than halving the resistor value 1k5 to 666, the diode voltage will only increase by ~0.4v?

that is correct.
But the real advantage lies in the lower dynamic resistance of the zener diode which falls rapidly reducing hum and supply nasties

The values don't have to be exact nor the same.
Just close to 18V, there is always a certain tolerance for each part.

[Deleted]

...... where do you start here? Just take a box of different value resistors and keep swapping them? Start at 0 Ohms and work up ....... start at 100 Ohms and work down? What kind of effects should be obvious? :-[CheersDicky

This is all uncharted territory

It should lower the internal resistance of the power supply.
Since the input stage is class A it is debateable what is audible or not.
In class AB it is evident.

try 100 Ohm and 0 Ohm and see/hear it's effect.

real pioneering

[Deleted]

Am I seeing a wascal wittle wabbit with a purple AD700 ... on it's hat ?

A w1000?

I loved your numerical reply!!! Talk about exactness? It just tickled the old rabbit tail.

Frans, the output resistance seems to be a balancing act. I have played around with your rather nifty device and there are differences in the balance with different resistance but there is a point where the output of some amps simply runs out of steam and it becomes lifeless.

So it seems that the output impedance is truly a balancing act between tone/impact. I kind of see that low output amps can be problematic if the output impedance is low and you're after the 'so called' flat response.

It's really hard to reach an ultimate conclusion since it's subtle and depends very much on what you're listening to.