Help needed to upgrade the capacitor in OTL tube amp
May 23, 2012 at 3:52 PM Thread Starter Post #1 of 113

telecaster

1000+ Head-Fier
Joined
Mar 7, 2008
Posts
1,059
Likes
365
Hello there! I need some help to upgrade the capacitor in my La figaro 339. Please correct me as I know next to nothing appart that there are lethal voltage all around.
 
The 4 little purple caps I think are the "lining" caps (don't know the name), they are 220µF 200V each.
The 6 bigger purple caps are the "power filter" caps I believe 820µF 200V each.
And there are 4 hidden "coupling" caps 63µF 350V each, I believe that they are bypassed by those rectangular blue caps. And if I swap them for MKP caps is it correct that I won't need to bypass them anymore with caps?
 

 
I want to increase the value of the coupling caps to 100µF each as I want more bass.
I choosed those MKP caps:

 
 
I want to change all the electrolytics caps to MKP but the 220µF ones are problematic because of size, so for them maybe I will just use Vishay BC electrolytic of a bigger value.
 
Is it good to increase the µF for the power filter and line capacitors?
 
Also using these MKP instead of the lytics 220µF is it going to upgrade the sound:?

 
I have so many questions, I hope someone will chime and help!
 
Cheers!
 
May 25, 2012 at 8:35 PM Post #2 of 113
I think you need to learn a lot more before you go trying to replace the caps with different types and sizes.  Unless I'm mistaken (I don't see a lot of transformers around), that thing is an OTL (Output Transformer-Less) amplifier.  As such, it needs some pretty huge caps at the output to keep from cutting out the bass due to the RC-cut-off frequency relationship.
 
In other words, all things being equal, you may need as much as 220uf at the output to keep the -3db cutoff at 20Hz with 32 ohm headphones.  Even then, the actual start of the cut in bass begins somewhere between 100Hz and 1000Hz.
 
So, what I'm saying is that attempting to replace electrolytic output caps with film capacitors at the output on an OTL amp is not going to work.  You simply can't purchase film caps with enough uf's to improve the bass.  What you should focus on is replacing the electrolytic caps with something even bigger.  If the coupling caps are truly 63uf, then I'd focus on finding some electrolytics as high as 220uf, 470uf, or even 1000uf, if you can afford them.  Bypass them with some good film caps and you won't notice the loss of quality with the electrolytics.  The film caps will carry most of the frequencies that are important.  Bass detail is not easily detectable, so you can make many cahnges while you're trying the increase the bass. 
 
May 25, 2012 at 10:21 PM Post #3 of 113
I use this OTL with HD650 which are 300 ohms, that gives that I need 265µF/channel (with a calculator it gives me 2Hz -3dB point and perfect 20Hz response point). I would need to buy two 130µ to put in parallel, it's not impossible but I'm searching if going all MKP and that's regular high voltage MKP not the fancy one is going to be better than electrolytic bypassed by high end cap...
The price in final is somewhere like 120€ for four 130µF MKP caps, or 60€ for good electrolytics and maybe 60/100€ for good bypass caps...
 
I still wonder if bypassing the four 130µF MKP is going to help to not...
 
 
May 25, 2012 at 11:52 PM Post #4 of 113
Well, OK - as long as you have a handle on how much uf's you need ... and it sounds like you do.  I haven't heard of anything this extreme in awhile (Ferrari did all film capacitors as the output capacitors once on a Millett Hybrid), but more power to you!  Bottom line, yes - if you can make up the uf's with film capacitors, they should almost always be better than even the best electrolytics. 
smily_headphones1.gif

 
May 26, 2012 at 12:05 AM Post #5 of 113
Those caps are huge and they fit just tight by my calculations!
 
Do you know if bypassing the big MKP will do any good, and if yes what value would be adequate?
I am searching for the resonnance frequency of various MKP cap and I can't find anything?! I thought I needed this info to choose the correct bypass cap value?!
 
 
May 26, 2012 at 7:57 AM Post #6 of 113
The whole reason for bypassing is to allow a film capacitor (sounds better, faster, more detailed) to take over from an electrolytic at the higher frequencies (greater than bass) that are more sensitive to human hearing.  If you don't have an electrolytic, there's nothing to bypass.
 
May 26, 2012 at 9:00 AM Post #7 of 113
Thanks for the info. If I understand this well, adding bypass capacitors will introduce distorsion (time smear?) at the crossover points? So a single capacitor handling the whole human sensitive frequency range without any crossover point is better? So the primary purpose of those low µF exotic and really expensive capacitors are for passive loudspeaker network? That would mean that bypassing capacitor is only a last resort solution to heal electrolytic problems?
 
May 26, 2012 at 10:27 AM Post #8 of 113
If you use only 300 Ohm phones, even one 130uf or 100uf (white one in your pic) make adequate bass, no need to parallel 2.
 
If you want upgrade PS caps, you can change only one el cap with film. I see, there is CRCRC filter, change only last C, closer to amp.
Easy, cost free PS upgrade - wen you change output caps to film, solder remaining blue film caps parallel to power caps.
 
May 26, 2012 at 10:52 AM Post #9 of 113
Quote:
Thanks for the info. If I understand this well, adding bypass capacitors will introduce distorsion (time smear?) at the crossover points? So a single capacitor handling the whole human sensitive frequency range without any crossover point is better? So the primary purpose of those low µF exotic and really expensive capacitors are for passive loudspeaker network? That would mean that bypassing capacitor is only a last resort solution to heal electrolytic problems?

Electrolytics have important applications with low-impedance headphones, too.  At 32 ohms, you need at least 1000uf to get to the same cutoff value that Zigis suggests above.  Also - IMHO, every electrolytic in the signal path should be bypassed.  The point being is that a film cap always sounds better, they're just not practical (or affordable) at the lower frequencies.  Bypassing is more of an art, too - and the size of the bypass is not really proportional to the electrolytic rating.  Usually, 0.47uf is about the limit of a good bypass, regardless of the electrolytic size.  I use 0.22uf the most, with 0.1uf being about the lower limit.  There are many schools of thought (smearing is one of them), but after much experimentation, I tend to think that a successful bypass combination results in a unique capacitance, not really two individual capacitors working on separate frequencies.
 
Here's a good read on output coupling capacitors and the results of several bypassing combinations:
Notes on Output Coupling Capacitors
 
May 26, 2012 at 11:25 AM Post #10 of 113
Quote:
If you use only 300 Ohm phones, even one 130uf or 100uf (white one in your pic) make adequate bass, no need to parallel 2.
 
If you want upgrade PS caps, you can change only one el cap with film. I see, there is CRCRC filter, change only last C, closer to amp.
Easy, cost free PS upgrade - wen you change output caps to film, solder remaining blue film caps parallel to power caps.


Thanks, I will bypass the power cap then with those blue film. To be sure, they are 0.47µF 250V.
 
And to be certain about the aforedmentionned cap, it is the cap that is the closest to the tubes? Also each of the power caps are 220µF in value, so it may not fit inside the case. Also costwise is it worthwhile considering I would spend also 120€ for changing only the both last power cap!
 
May 26, 2012 at 11:38 AM Post #11 of 113
Quote:
 Usually, 0.47uf is about the limit of a good bypass, regardless of the electrolytic size.  I use 0.22uf the most, with 0.1uf being about the lower limit.

Thanks a lot! One thing I still wonder, those bypassing capacitor value is for a single cap and should not be the total of all the bypass caps in the row of paralleled electrolytic caps?
 
May 26, 2012 at 11:45 AM Post #12 of 113
I don't know if it's worthwhile to increase the imput capacitor value or just replacing them with two 22µF film cap. I guess I don't know the imput impedance so I cannot calculate de frequency cut off.
 
May 26, 2012 at 3:20 PM Post #13 of 113
Quote:
I don't know if it's worthwhile to increase the imput capacitor value or just replacing them with two 22µF film cap. I guess I don't know the imput impedance so I cannot calculate de frequency cut off.

Typically, the input impedance is set by the impedance of the volume pot.
 
May 26, 2012 at 6:06 PM Post #14 of 113
the engineering numbers make it hard to believe in bypassing audio signal coupling electros with few % film/mica/PIO, whatever FOTM
 
bipolar - full depth oxide on both foils is "best" on measured distortion for electrolytics - if you care you just use brick sized film caps alone - they mesure better than any electro
 
I have measured a multichannel SACD player and found lower distortion with the "ordinary" Nichicon on the surround channels vs the Muse caps used in the R,L channels - the distortion was only just visible on the high pass "shoulder"
 
the condition for distortion from cap dielectric is that there has to be signal drop across the cap - at filter edges - so just "supersize - make the coupling C 10-20x larger that "needed" and the audio distortion is impossible to measure
 
the numbers for a small bypass cap shorting out a decent electro's ESR is laughable also - the corner frequency is typically 10s of MHz - no effect at audio - and there is a typically L:C resonant peak from lead inductance at lower frequency
 
I realize there is huge audiophile belief in "Capacitor Sound" - but no "Capacitor Tasting" review I've ever seen goes into any depth about even basic Blinding, controls needed to make hearing the "only" thing used for the discrimination 
 
May 26, 2012 at 6:28 PM Post #15 of 113
Here's an excellent article on "Blinding":
http://www.innerfidelity.com/content/beyond-neutral-measurements-blind-testing-subjective-experience-and-personal-pleasure-headph
Just someone's (
wink.gif
) opinion, but it's interesting to note this quote: "I do not think it [blind testing] does a very good job of quantifying performance characteristics, because humans are not at all like scientific instrumentation."
 
I would also note that bypassing output coupling capacitors is not a FOTM fascination and has been used since well before 2000 for headphone amplifiers.  Besides Dsavitsk, Pete Millett has used bypass capacitors in many of his designs, along with Alex Cavalli, Andrea Ciuffoli, and van Waarde.  Amps are designed with either sand or glass, or a combination of both.  Those that specialize in glass and combinations with sand are very familiar with bypass capacitors.  The all-sand proponents typically don't have a clue. 
 

Users who are viewing this thread

Back
Top