[Reverie]

Hey all,

I’m coming to the last stages of my Amp4 project, where I’m deciding on some of the perhaps more trivial parts. I’ve already sorted out the input caps (Jensen 2.2uf PIO), output caps (Wima MKP2) and torioid inductors (Metglas Cores).  However, there are just a few bits I’m undecided on, and would appreciate any feedback available.

Firstly, the electrolytic caps. While many claim Blackgate are the undisputed champion here, they are indeed a considerable amount of money (that said, I did get Jensen PIO for the input ) and more importantly, I trust Michael Mardis’s word when he says they aren’t really suited for this amp. Therefore, after considerable searching, I’m left with generally two choices that seem good for most caps. Firstly, the Elna Cerafine. From what I’ve read about them they offer a good compromise between the ‘naturalness’ of the Silmic II, and the  ‘directness’ of the Blackgates, and I’m assuming they’re suited to this application. My second choice is the Panasonic FM series. While rather less impressive, they’re still meant to be ideal for this situation, and are rather cheap too.  The various capacitor situations and my thoughts on it are as follows:

2x 10000uF 50V – These caps I do not need to consider, as they’re for AC smoothing, which is unneeded with my battery source.

1x 10uF 10v , 1x 1000uF 10v, 2x 330uF 35V – These caps can all be Elna Cerafine quite easily. There are several places that sell this size range of them but Audio-cube (although they’re currently on holiday) have the best range.

2x 1500uF 35V – These power supply caps are rather important. However, I’ve found it quite hard to find Cerafines in a suitable voltage/size.  Audiocube doesn’t stock them, and while the PDF they link to on that site doesn’t say that Elna do anything that large (might be old) I’ve seen several 10KuF caps and some 63V ones on ebay in the past, so they must have been made sometime. Anyone any ideas on where I might be able to source these? If all else fails, I could quite happily go for the Panasonics here, probably to be sourced from Farnell UK.


Next issue, is the input hook-up wire. I’ve been convinced using shielded wire is the way forward, as I have a particular noise issue in my amp’s to-be home (PC, and various other noise makers). However, there’s still quite a list that I have to choose from. If you’d like to express any opinions on your preferred it would be great.

Weisd - Belden 8442 $1.02 / foot
DIY Audio Store- Oyaide Pa-02 $35 AUS /metre ($27.57USD) (A bit pricey...)
DIY Audio Store- Mogami W2893 $4.25 AUS/metre ($3.35USD) – AWG26
Percy Audio - Mogami 2534 - Four conductor (24ga) 6mm D shielded microphone cable OFC Neglex copper $1.15 per foot AWG24
Percy Audio - MUSIC METRE IV A quasi solid core design comprised of four insulated strands, 3x28ga & 1x34ga per + & -, utilizing a high purity Teflon insulated OF copper, foil shield (40pF foot shielded),  7mm D… $7.50 mono foot
Percy Audio - Canare Star Quad: The L4E6S four conductor (21ga) 6mm D microphone cable 44, 57pF/foot, black 6mm D… 95¢


Next, I’m considering making myself some new Cat5e speaker cable, specifically VHaudio’s DIY project. However, the site does state that they will be of high capacitance, and that this could prove a problem for some amps. Even though I’m not going to be using lengths any long than say 2.5m (and there are Zobel caps in the amp4 design), would this be the case with Tripath amps? If so, what would you advise? 

Lastly, I’ve already chosen my attenuator from the bargain one on diyfidelity, but I’m still undecided as to which resistance is best. I remember someone telling me Scott Endler said 10Kohm was preferred, while Vinnie of Red Wine Audio fame prefers 20kohm. I’m not really sure how to judge this.

Oh, and on a side note – banana plugs…what are your favourite?

If anyone can offer any advice to help me finish off this project, it would be greatly appreciated. 

Thanks to all who post,
Peter 

[Scotty]

I would go for the FM caps they are available in the sizes you mentioned except for the 10mfd which you could possibly replace with a 33mfd at16v,from Digi-Key.See link:   http://www.digikey.com/scripts/DkSearch/dksus.dll?Selection
I have used FM's extensively as coupling caps and as local power supply storage caps. I would also re-think those 10,000mfd storage caps. A pair of 10,000mfd Panasonic FC caps doesn't cost much.
A good pair of low ESR caps of that size may yield a substantial improvement in
the dynamics and bass impact the amp has because they can have the ability
to store enough energy to power the circuit and deliver it more quickly than the battery. The ideal is to have a power supply with a low dynamic impedance at audio frequencies. I would not make assumptions about what the battery does on a dynamic basis.
For my Tripath amp, I used a 12ga. multi-strand twisted pair that I shielded with copper foil grounded at the input jacks for my input wiring.
I would use as large a gauge wire as I could reliably solder to the board for the output wiring and keep it as short as possible. The length of the power supply wiring you use is even more critical to avoid power supply ringing from the the inductance in the wiring from the supply to the board.
Bear in mind that no matter what you do regarding your choice of wires or capacitors you cannot build a perfect power amp, and the chances are that the power amp you are building has already been made obsolete by something else that you don't know about. In short don't obsess over the parts selection process,focus on building a working amplifier and enjoy the musical results.
Scotty

[Reverie]

I would go for the FM caps they are available in the sizes you mentioned except for the 10mfd which you could possibly replace with a 33mfd at16v,from Digi-Key.See link:   http://www.digikey.com/scripts/DkSearch/dksus.dll?Selection
I have used FM's extensively as coupling caps and as local power supply storage caps. I would also re-think those 10,000mfd storage caps. A pair of 10,000mfd Panasonic FC caps doesn't cost much.
A good pair of low ESR caps of that size may yield a substantial improvement in
the dynamics and bass impact the amp has because they can have the ability
to store enough energy to power the circuit and deliver it more quickly than the battery. The ideal is to have a power supply with a low dynamic impedance at audio frequencies. I would not make assumptions about what the battery does on a dynamic basis.
For my Tripath amp, I used a 12ga. multi-strand twisted pair that I shielded with copper foil grounded at the input jacks for my input wiring.
I would use as large a gauge wire as I could reliably solder to the board for the output wiring and keep it as short as possible. The length of the power supply wiring you use is even more critical to avoid power supply ringing from the the inductance in the wiring from the supply to the board.
Bear in mind that no matter what you do regarding your choice of wires or capacitors you cannot build a perfect power amp, and the chances are that the power amp you are building has already been made obsolete by something else that you don't know about. In short don't obsess over the parts selection process,focus on building a working amplifier and enjoy the musical results.
Scotty

Hi Scotty,

Thanks for your wise words. 

The FM caps are indeed very cheap, so I'll probably buy them anyway, as the cost is negligable compared to the cost of the project overall. While I'm still keen on the cerafines, as you speak so highly of the FMs, I shall only use the cerafines where I can get the exact values, and not subsitute in similar with the ceras, but use the correct FM instead. I probably would not have done it that way otherwise. 

However, I would like to take you up on your point about the 10,000uF caps. To explain exactly what ones I'm talking about, I've made a slight alteration of the diagram from part of the schematic (illistrating the various parts, and taking out irrelevant values to protect Jan's design):



Now the 10,000uF caps are the ones in the AC section and, as far as I am aware, perform the function of 'smoothing'. As I'm planning on missing this entire section out and using DC, I'm pretty sure I won't need these at all. However, I may have confused you into thinking these were the power supply caps (I'm still learning a lot about amps and circuits in general), because I was under the impression the role you describe is in fact that of those 2*1500uF caps (green section on diagram).

Also, I thought that a large gauge of wire was only favorable for output wires, and that instead, a gauge of about 23AWG was better. However, I never quite understood the reasons for this, so I cannot proove it.

Thanks again!

Peter

[Scotty]

Peter,I am not sure where the power supply stiffing caps are located in regards to the entire circuit. I would use a FM cap for C1, substituting a 1000mfd 35v FM. If the 1500mfd are on the amp circuit board I would for sure use the FM caps instead of the Cerafines,the FM caps are designed for switching applications which is what your amp is. How I would do this project assuming that when you run off the battery you are disconnected from the AC mains, is to put the 10,000mfd
caps in parallel with the battery supply. I would put one 10,000 mfd cap across each battery supplying the positive and negative rail voltage. Your charging
power supply only has to provide a positive charging voltage to the batteries
and you would connect the caps to the batteries observing the marked polarity on the caps connecting negative as marked on the cap to ground as marked on the battery. The negative rail voltage is derived from reversing the battery polarity terminal hookup to the board. The batteries take the place of the mains transformer keeping the caps topped off with the voltage and current to run the circuit. Like I said the FC caps are cheap and you can
do a test to see if you can hear an improvement with them in the circuit or not. If you wire a single pole single throw switch between the cap and the supply rail it's connected to you could play the amp with them out of the circuit with switches open then power down the amp and close the switches
and play it with the caps in the circuit and see if you hear an improvement.
If I was building a Hotrod version I would use 10,000mfd Jensen four pole caps in place of the FC's as they have a very low impedance up to 1mHz and were designed for high speed switching power supplies. You Tripath may switch at speeds up to 1mHz.
Scotty

[Reverie]

Wow  , thanks for the detailed explanation. 

Looks like I will not be using any Cerafines due to their lack of suitability; FMs all the way! (might have a chance to try a pair of Jensesn Elecs too if I get the chance - they're not too expensive). Your idea for testing the difference made by the mod sounds very promising too.

If I have any questions about your idea I'll get back to you .

By the way, I forgot to ask - in your first post where you mentioned the power supply, were you suggesting shielded hook-up was the best way?

[Scotty]

Some sort of shielded hook up wire on the power supply from the caps/batteries
to the amplifier board would be a good idea. A very short run of shielded 14ga. wire might keep the radiated RF noise down inside the chassis and the capacitance per foot would tend to offset the inductance unshielded hookup
wire has which could help reduce power supply ringing. The Jensen 4pole caps run about $38.50 a piece and up, from RAM: see link  http://www.referenceaudiomods.com/Merchant2/merchant.mvc?Screen=CTGY&Store_Code=RAM&Category_Code=JE4PA
Not exactly cheap but the best for this kind of application.
Scotty
Here is a link to the history of the 4 pole electrolytic capacitor.
http://www.jensencapacitors.com/audio/white-papers/4pole.html

[Gordy]

To back up what Scotty is suggesting for your power supply caps, before finalizing the  battery supplies for my UcD 400 amps I did extensive listening tests with varing amounts of capacitance.  I used both my chips amps and the UcD's actually and tried batteries alone, 500uF, 1500uF, 4.7k and up to 15K finally settling on two 4.7K caps as my preferred balance of speed and dynamics.  My belief is that using multiple smaller caps (all bypassed as well), as opposed to a single 10k cap, is the way to go. Remember that one advantage of battery supplies are thier basically zero impedience so, the more caps the merrier?  I'd be very tempted to go with two bypassed 3k uF panasonics followed by a 3k 4 pole! YMMV of course    An inexpensive and very good bypass cap is the Jantzen poly cap... http://www.partsexpress.com/pe/pshowdetl.cfm?&PartNumber=027-910&DID=7

[Scotty]

Listening tests have shown that the presence of caps that have a higher inductance than the Jensen 4poles in when used in parallel with them actually degrades their performance. Go figure, it seems that it is an all or nothing proposition with the Jensens.
Scotty

[Gordy]

Thanks Scotty!  Jensen's site actually rec'd the idea (Jensen following the std. caps) and I was going to give it a try in another ps.

[Reverie]

I can see your thought behind this idea. I'll draw a diagram soon to check I've got the layout correct.   

I too had heard what Scotty said, about the loss of performance of high quality caps if pairred with lower quality ones.

I would like to know your thoughts on this battery-parallel capacitor's importance though - Is there a case for using the Jensen 4 pole? While they're not *too* expensive, I'd rather not pay all that for a pair if I'm not going to come close to hearing any difference.

Thanks for your time Gordy and Scotty

EDIT: Just looking...I can't actually find any FC or FM for 10000uF 35V, only 10v .

I also had to substitute the 10uf for a 100uf FM due to what Farnell stocked...I didn't think it was particularly value-specific   

EDIT 2: If it is value-specific, I could use a 10uF 35v FC, thoughts as to which is better?

[ohenry]

I, too, have found the Panasonic FM caps to be very good in T-amp applications.

I used unshielded 22 ga. UPOCC solid copper for the input signal. I liked the results in the amp4 and had no noise issues.  For grins, I recently replaced the UPOCC input wire with shielded two-conductor 28 ga. Belden and the result was pretty dreadful, don't go there.  If shielding is desired for the input wiring, you may want to insert the premium wire of your choice into a metallic mesh.

Good luck and have fun.

[Reverie]

I, too, have found the Panasonic FM caps to be very good in T-amp applications.

I used unshielded 22 ga. UPOCC solid copper for the input signal. I liked the results in the amp4 and had no noise issues.  For grins, I recently replaced the UPOCC input wire with shielded two-conductor 28 ga. Belden and the result was pretty dreadful, don't go there.  If shielding is desired for the input wiring, you may want to insert the premium wire of your choice into a metallic mesh.

Good luck and have fun.

Hmm, your experience with the input wires is both interesting and worrying Henry.

I've currently drawn it down to the Mogwai and the Music Metre, hoping to order both and decide at the time. I probably should add some OCC FEP-insulated copper too though, based on your results.  May I ask where you purchased yours from? If you have any spare, I'd be happy to buy it off you  .

That reminds me, I need to reply to you about the Electrum Audio wire, there's been a bit of a development... 

[Gordy]

I'd love to give the Jensens a try in a std. PS some day, Wayne at Bolder Cable uses them in his top of the line PS's for the Squeeze Boxes etc. 

Here's a very good 10k uF 63v cap... Panasonic TSHA

[ohenry]

Hi Reverie,
I would hope that your shielded wire is better quality than mine.  I bought a few feet of cheap surplus and it's quality is nothing special, so don't run away from your selections.

I think I got the copper from Parts Connexion, or it may have been left over from some that I ordered from Audiocom.com (from your side of the pond).  Unfortunately, I don't have any left.

I recently bought a few feet of cryogenically treated Oyaide PA-02 to make some interconnects, something similar to that may be good.  I haven't tried it yet, so I can't recommend.  You may want to take a look at Cryo-Parts.com to see if there is anything that interests you.

[Reverie]

Here's a very good 10k uF 63v cap... Panasonic TSHA

Looks good, but what voltage is needed? Surely 35v would be fine here, considering each battery they will be in parallel with is only 12v...

[Gordy]

Sorry, that just happened to be what I was using in a supply, try searching Digi-key part # P6573-ND for a 25v version...

[Reverie]

Here's my attempt at drawing the power stage of the amp, as promised.



I'll reply to the other posts in a minute.

[Reverie]

Sorry, that just happened to be what I was using in a supply, try searching Digi-key part # P6573-ND for a 25v version...

Thanks, that cap looks good! It's ordered for that place.

I bought a few feet of cheap surplus and it's quality is nothing special, so don't run away from your selections.

Ah that's fair enough, I'll stick to my guns.

I think I got the copper from Parts Connexion, or it may have been left over from some that I ordered from Audiocom.com (from your side of the pond).  Unfortunately, I don't have any left. Sad

Ah, that might be the Neotech. Looks like good stuff, the only thing that put me off was the PVC dialelectric. Then again, I don't *actually* know how important that is.

I think I'll get a few feet of Vampire CCC 2x 18AWG interconnect wire from Percy Audio. Although it's a dual conductor arangement (I think that's still *kinda* solid core...), it uses polyethylene insulation and has a copper shield.

I recently bought a few feet of cryogenically treated Oyaide PA-02 to make some interconnects, something similar to that may be good.  I haven't tried it yet, so I can't recommend.  You may want to take a look at Cryo-Parts.com to see if there is anything that interests you.

That site looks pretty good. Are there any minimum order requirements?

I've heard about the PA-02, supposedly legendary. ~$20/metre isn't cheap though, though I imagine it could turn out very good value...

By the way, please feel free to comment on my drawing. It's probable I've made some mistakes.