Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Help needed to upgrade the capacitor in OTL tube amp
New Posts  All Forums:Forum Nav:

Help needed to upgrade the capacitor in OTL tube amp - Page 5

post #61 of 113
Quote:
Originally Posted by telecaster View Post

That was close I was almost beginning to forfeit!

Anyway Wall-E what sort of resistor are you gonna use as output coupling resistors? The Mills don't seem to handle a lot of watt... I figured out my SETA are inductive so it might be a problem, what do you think?
 

I checked voltage and current consumption on the cathode resistor and depends of what tubes I used it is from 84-90V and current is 168mA-184mA what give the Rk wattage between 14-17Watt idle and with the load on is almost the same. Of course better oversize the resistor to hold in reserve few watts or use smaler in parallel which will disperse the heat better then single one.The Rk like output coupling caps have an effect on sound and better use good components.
I'll use non inductive 1% Wirewound resistors like, Mills  MRC50 this one is 50Watt or go with MRA-12 series 12Watt so 3 in parrarel give me 36watt seems enough.

post #62 of 113
Thread Starter 

Ok thanks, I have received the solder strip but unfortunately it's a small model, not like the one Yuking is using. anyway I use my seta resistor and it fits the amp, I will eventually replace the whole array when I spot a problem, but for now the sound is pretty much problemless even though those resistors are inductive...

 

I added 30uF per channel of MKP totallizing 230uF per channel of film cap. The bass is slightly better now compared to stock, I have spotted the sound is more well rendered in the lower octave, and in general bass has more body,but it's slight.

 

The sound of the MKP vs lytics is also a slight change but for the better. I don't bypass the flim caps at all, I find the sound detailed and more relax and in comparaison the lytics were a little scratchier in the top end, and the bass lacked the lowest octave probably due to the 130uf per channel value of stock lytics.

The sound is more coherent and has more depth with the film cap, but it's not a night and day change, as I did the switch I know it's there, but it's not "obvious".

 

I added 4.7uF bypass to the 470uF lytics in PS, and 9uF bypass cap to the last 820uF cap in the PS. I found the sound a little cleaner, but the change is really slight too, but I like it better that way. Costwise those are really not worth it, but in a cost non object I would say why not...
 


Edited by telecaster - 6/8/12 at 8:26am
post #63 of 113
Thread Starter 

Added 10000pF silver mica to the output, still testing, the sound is not bad...

The 4X100µF MKP are on the other side of the board. Still waiting for the PS caps to arrive..

Gotta find a way to fit two more 30µF MKP in there somewhere!

 

_DSC2360-2.jpg


Edited by telecaster - 6/9/12 at 3:03am
post #64 of 113

I'm glad you're happy with the results you did to modifying your lf339 anyway pleased to Do It Yourself.


Also I would not expect huge improvement like Night & Day, look at this amp is build like tank every thing is doubled and  the performance with low-high cans & SQ are first class. It's not like my 332c which looks poor compared to lf339  in which I upgraded output caps and the result exceeded all my expectations I could tell that it was almost like Night & Day.
But for me even small improvement will be worth to put the extra DIY wchich i like it and "£,$,€" to tweaking the amp.


Good Job!!! telecaster.

Quote:
Originally Posted by telecaster View Post

Ok thanks, I have received the solder strip but unfortunately it's a small model, not like the one Yuking is using. anyway I use my seta resistor and it fits the amp, I will eventually replace the whole array when I spot a problem, but for now the sound is pretty much problemless even though those resistors are inductive...

 

I added 30uF per channel of MKP totallizing 230uF per channel of film cap. The bass is slightly better now compared to stock, I have spotted the sound is more well rendered in the lower octave, and in general bass has more body,but it's slight.

 

The sound of the MKP vs lytics is also a slight change but for the better. I don't bypass the flim caps at all, I find the sound detailed and more relax and in comparaison the lytics were a little scratchier in the top end, and the bass lacked the lowest octave probably due to the 130uf per channel value of stock lytics.

The sound is more coherent and has more depth with the film cap, but it's not a night and day change, as I did the switch I know it's there, but it's not "obvious".

 

I added 4.7uF bypass to the 470uF lytics in PS, and 9uF bypass cap to the last 820uF cap in the PS. I found the sound a little cleaner, but the change is really slight too, but I like it better that way. Costwise those are really not worth it, but in a cost non object I would say why not...
 


Edited by WALL-E - 6/9/12 at 4:41am
post #65 of 113
Thread Starter 

Thanks for the kind word!

I don't know if if it is a placebo effect but the sound is really better than before all the mods! The 339 highs were already very clear and relax, but now it's another level of clearness and naturalness. It is like a layer of mild distorsion has disapeard. Now there are 3X 10000pF of silver mica per channel, it's the russian capacitor CRM3.

I think now the sound is even more clear in the upper range. I will stick with that for some days to analyse the effect of those bypasses.
 

post #66 of 113
Thread Starter 

Some days have passed and I can tell now that the bass is definitely better! I suspect it is the 470µF caps in place of the 220µF.

 

Before the mods the highs were more agressive and the bass less warm and less powerfull. Now The highs are perfect mix of detail and relax, and the bass has more power and more definition. This is surely due to the  film output coupling caps that is almost double of stock value. The silver mica cap also help opening the highs that seem more natural than without it. But this is not certain I need more time to tell the differences.

 

That is my findings so far!

I have ordered  some non inductive caddock to replace them in CRCRC filter and and output coupling. Also have some carbon composition to try at pin 4 and pin 5 in place of the white box...


Edited by telecaster - 6/12/12 at 2:18pm
post #67 of 113
Thread Starter 
Quote:
Originally Posted by WALL-E View Post

I'm glad you're happy with the results you did to modifying your lf339 anyway pleased to Do It Yourself.


Also I would not expect huge improvement like Night & Day, look at this amp is build like tank every thing is doubled and  the performance with low-high cans & SQ are first class. It's not like my 332c which looks poor compared to lf339  in which I upgraded output caps and the result exceeded all my expectations I could tell that it was almost like Night & Day.
But for me even small improvement will be worth to put the extra DIY wchich i like it and "£,$,€" to tweaking the amp.


Good Job!!! telecaster.


Is this rectifier a correct replacement?

GI2502 Rectifier bridge 200V 25A

post #68 of 113
Quote:
Originally Posted by telecaster View Post


Is this rectifier a correct replacement?

GI2502 Rectifier bridge 200V 25A


25A bridge is way too much 3A could be enough for the headphone amp but I would not worry it's doesn't matter but max voltage does matter. In theory 200V bridge could work with the 140V AC from the secondary winding of transformer but in practice could be failure. Beacouse AC from the outlets are not constans and not clear of noise the voltage can jump for fraction of a microsecond and the peak impuls (overvoltages) could  damage one of diodes in the bridge in the best case the fuse blow up but if not all the electrolytic capacitors overheats then get swollen and....Bum!  Ok all bridges over 200V seems enough.

May I suggest  you the bridges/diodes use in critical audio applications like HEXFRED high-speed soft-recovery Bridge Rectifiers.
IXYS FBE 22-06N1  22A 600V
IXYS VBE 17-06NO7 27A 600V

or you can build Gretz bridge with four BY228 (1500V 3A) the same as in keeph modded LF339 and very cheep.

or the best of all Schottky bridge rectifiers IXYS SiC but  they are very exspensive!!! and probably not worth it.


Edited by WALL-E - 6/13/12 at 2:32pm
post #69 of 113
Thread Starter 
Quote:
Originally Posted by WALL-E View Post

May I suggest  you the bridges/diodes use in critical audio applications like HEXFRED high-speed soft-recovery Bridge Rectifiers.
IXYS FBE 22-06N1  22A 600V
IXYS VBE 17-06NO7 27A 600V

or you can build Gretz bridge with four BY228 (1500V 3A) the same as in keeph modded LF339 and very cheep.

or the best of all Schottky bridge rectifiers IXYS SiC but  they are very exspensive!!! and probably not worth it.

 

Thanks for the advice! It really helped me. I have looked at the IXYS HEXFRED rectifiers and "IXYS VBE 17-06NO7 27A 600V" looks like it's not soft recovery (which seems to be the best for audio). And "IXYS FBE 22-06N1" is soft reverse recovery, don't know if it's the same as soft recovery.

 

For the best SIC rectifier, those are 80 bux a unit, and I have the idea to make a cheaper bridge with SIC schottky using those : http://uk.farnell.com/jsp/search/productdetail.jsp?sku=1851840

What do you think? It seems it would be the same quality using SIC schottky diode to make a bridge from scratch?

 

I have read some terrible things about diode that blow up, are you sure if I use 4A sic schottky to make a bridge it is safe enough?

 

Also is we use those fast schottky, will it be safe using 1000µF 200V right after it? I have read that using fast diodes increased the voltage in the PS filter...

 

Also I have experimented replacing the white box 1K ohm with allen bradly carbon comp resistor, and I found the sound slightly inferior to the white box. But with the carbon comp I have no hum when using EF80, but I have hum in one channel using my tungsol 6SJ7GT. But using the white box I have no hum using the TS 6SJ7GT and incredible hum using EF80 in one channel... Pretty weird isn't it!

I was unable to open this white box because it is glued.

 

 


Edited by telecaster - 6/14/12 at 3:18am
post #70 of 113

HexFred diodes/bridges are fast switching but soft response,which means it has less switching(ringing) noise to provide a cleaner, faster DC voltage. The HexFred have good opinion and is often used as a replacment of standard rectifiers in audio hi-fi / hi-end power supply.
The main difference between HexFred and Schottky is the Schottky are low forward voltage drop and no reverse recovery. So bulding the schottky diodes bridge will increase the voltage but don't worry just slightly. The general purpose single rectifier diodes will forward voltage drop about 0.7V and for Schottkthy is 0.2V.
Transformer in the power supply give 140V with the max current about 300mA and 6,3V/3A for filament which is supply directly from AC transformer so the bridge will use only with high voltage so the 4A sic schottky are safe enoug.
You can build the Graetz bridge and you will get the same parameters like singe bridge rectifiers.
 


Edited by WALL-E - 6/14/12 at 5:05pm
post #71 of 113
Thread Starter 

Thanks for the reply m8! I have finally decided on Vishay HFA08TB60, 8A 600V hexfred because I have read some people finding the SIC kinda hard sounding... Which is something I don't like in general, and also because farnell wanted 20€ shipping cost for the SIC and the only ebay seller who has them don't ship abroad.

 

I have made a little graetz on a small heatsink that will fit nicely, here is a picture of my schematic before powering it because I don't want to make everything explode! I have found those diodes cheap hence my choice over the IXYS FBE2206N1.

 

I have ordered four per channel 30W // caddock resistors in TO220 format to use as output resistors. They are non inductive and I hope will be reliable, less hot and good sounding. The seta works fine but I just want all the resistors to be non inductive and have a foolproof design. The PS resistors will also be change for caddock TO220 format.

 

Do you know what is the use of the 10K near the headphone jack output? I hear using carbon composition of high value and in bad places will produce distorsion. Is this 10K replaceable with carbon comp without drawback? Also can I replace the 1M resistors across the driver tube with carbon comp?

 

Thanks for all your help WALL-E, this is greatly appreciated!

 

Can you please tell me if this is correct?:

_DSC2383.jpg

 


Edited by telecaster - 6/14/12 at 6:41pm
post #72 of 113

The two 10k Ohm resistor next to headphone socket are bleeder resistor.The output of coupling capacitor is connected through a resistor to ground in order to prevent headphones from DC stored in capacitors, and the amp does have a load even with no phones connected so these two resistors bleed away the DC if no phones are connected. The 10k is a load like headphones but with 10 times more(resistance) than the higher impedance headphones I know. Probably not have a significant impact on sound quality but they are in signal patch. I would be willing to slightly increase bleeder resistor on the output something in between 10K & 22k Ohm. But increasing the resistance will increase the time to bleed away the DC.

Formula Tsec = R Ohms * C farads
some examples:(stock lf 339) 90V, 130uF with 10K Ohm bleeder resistors the time is about 1.30 sec aftre you can conect your hedphone without pop up sound noise in your hedphones.
90V, 260uF with 10K Ohm the time is about 2.60 sec
90V, 260uF with 22K Ohm time is about  5.72 sec and etc.

Your drawings of bridges are correct but if you not sure or just make a mistake u dont know I sugest to test the bridge with some small transformel low voltage like 230/12V  taking out from broken toy or something and check the bridge with mulimeter before you put it in amplifier.

 

The caddock like a Millet are on my top list, with TO220 housing caddock you will need heatsink to disperse the heat.

Quote:
Originally Posted by telecaster View Post

Thanks for the reply m8! I have finally decided on Vishay HFA08TB60, 8A 600V hexfred because I have read some people finding the SIC kinda hard sounding... Which is something I don't like in general, and also because farnell wanted 20€ shipping cost for the SIC and the only ebay seller who has them don't ship abroad.

 

I have made a little graetz on a small heatsink that will fit nicely, here is a picture of my schematic before powering it because I don't want to make everything explode! I have found those diodes cheap hence my choice over the IXYS FBE2206N1.

 

I have ordered four per channel 30W // caddock resistors in TO220 format to use as output resistors. They are non inductive and I hope will be reliable, less hot and good sounding. The seta works fine but I just want all the resistors to be non inductive and have a foolproof design. The PS resistors will also be change for caddock TO220 format.

 

Do you know what is the use of the 10K near the headphone jack output? I hear using carbon composition of high value and in bad places will produce distorsion. Is this 10K replaceable with carbon comp without drawback? Also can I replace the 1M resistors across the driver tube with carbon comp?

 

Thanks for all your help WALL-E, this is greatly appreciated!

 

Can you please tell me if this is correct?:

_DSC2383.jpg

 


Edited by WALL-E - 6/15/12 at 10:22am
post #73 of 113
Thread Starter 
Quote:
Originally Posted by WALL-E View Post

The two 10k Ohm resistor next to headphone socket are bleeder resistor.The output of coupling capacitor is connected through a resistor to ground in order to prevent headphones from DC stored in capacitors, and the amp does have a load even with no phones connected so these two resistors bleed away the DC if no phones are connected. The 10k is a load like headphones but with 10 times more(resistance) than the higher impedance headphones I know. Probably not have a significant impact on sound quality but they are in signal patch. I would be willing to slightly increase bleeder resistor on the output something in between 10K & 22k Ohm. But increasing the resistance will increase the time to bleed away the DC.

Formula Tsec = R Ohms * C farads
some examples:(stock lf 339) 90V, 130uF with 10K Ohm bleeder resistors the time is about 1.30 sec aftre you can conect your hedphone without pop up sound noise in your hedphones.
90V, 260uF with 10K Ohm the time is about 2.60 sec
90V, 260uF with 22K Ohm time is about  5.72 sec and etc.

Your drawings of bridges are correct but if you not sure or just make a mistake u dont know I sugest to test the bridge with some small transformel low voltage like 230/12V  taking out from broken toy or something and check the bridge with mulimeter before you put it in amplifier.

 

The caddock like a Millet are on my top list, with TO220 housing caddock you will need heatsink to disperse the heat.


Thanks a lot! I now understand the purpose of the bleeder resistor! But why would you want a higher value? Is the higher value safer and shunting more DC to the ground?

So if this bleeder is in the signal path, I guess using carbon composition would introduce distorsion? I thought that only carbon comp resistance in series would be introducing distorsion.

And as for the Graetz I'll make, that is a good advice to try on a low voltage transformer beforehand connecting to my amp!

 

For the output coupling resistors, will you choose the 50W Mills wirewound? Or the caddock? Frankly the caddock seem expensive at this position but I did it anyway, it would look original at least. And yes I have those heatsink on order for caddock TO220, now I have to think hard how to arrange all those components in this little 339 case!

post #74 of 113

The main reason to use bleeder resistor is safety and the higher value is not safer and will need more time to discharge cap but safer not always mean "better" in audio. The audio signal flows from input to the output and it will  “hit”  the bleeder resistor too. Every commponent got some noise in that case bleeder resistor which could have impact on sound quality. It’s too risky to remove the Bleeder at all but increase the value will make the “hit” less affected.

With the carbon resistors is like with the VDH carbon cables is known as warm and musical some people like them some dislike.
My advise is to use metal film as much as you can.

The stock 470ohm aluminum-housed resistor are bad they tend to drift in value with temperature, and time.
For a cathode resistor I will choose Wirewound  Mills MRC50 or MRA-12 both are non inductive and  1% tolerance in PS i will use caddock TO220 and elsewhere precision metal film resistors.

 


Edited by WALL-E - 6/16/12 at 8:58am
post #75 of 113
Thread Starter 

I haven't found 22K in my stach, could I use 25K without detrimental effect? It is very interesting thanks for the knowledge WALL-E!

 

I received the BC 1000µF 200V and installed them. I added 30µF MKP caps to the output coupling, and I removed all the bypass caps too. I think it must be the brand new caps but I find the sound lacking bass. The sound is relax though but the bass quality is not there. I will let the amp burn couple dozen of hours and see if it improves.

Overall I prefer the sound before this mod, highs were more calm, and bass were deeper stronger. i strongly suspect the caps needing burn in.

_DSC2391.jpg

New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Help needed to upgrade the capacitor in OTL tube amp