| toque_101 |
Building a GC with 2 channels (OPA548). I'm wondering what use for the input cap. I would go with black gate but I can't find anywhere in canada to get them easily. I have two options in mind that I have to decide on: a 4.7uF Solen Metallized Polypropylene Fast Cap, or a 4.7uF Bennic Metallized Polyester Cap. Any preference here or should I just close my eyes and point? :)
Also, does anyone know a good source for Welwyn resistors or just good GC resistors in canada?
Thanks! |
|
|
| Nuuk |
Polypropylene is a better choice than polysester.
In the UK, we can buy stuff from Newark through Farnell and I guess that it is a reciprocal arrangement. Also, look for Dale resistors which I believe are the same as the Welwyns. |
|
|
| peranders |
| A "cheap" advice: Take some polyester types from Wima, Roederstein (ERO) 1 uF/63V and if you are _very_ dissatisfied change to an another type. Small 63 V polyester is surpringly good, not so cool or exotic though. :no: |
|
|
| Nuuk |
| quote: | | Take some polyester types from Wima, Roederstein (ERO) 1 uF/63V |
Per, does that apply just for those smaller values or would a 4.7uF polyester be good too? (It was the input cap which was originally asked about here)
I have always gone for polypropylene for input caps but am open to try anything, especially if it is cheaper. ;) |
|
|
| jean-paul |
| Quality has a price. |
|
|
| peranders |
| quote: | Originally posted by Nuuk
Per, does that apply just for those smaller values or would a 4.7uF polyester be good too? (It was the input cap which was originally asked about here)
I have always gone for polypropylene for input caps but am open to try anything, especially if it is cheaper. ;) |
The smallness itself is good if we talk HF properties. I haven't seen any bigger than 1 uF/63V.
I like polypropulene better but this is (I think) because they are electrically better.
Quality, what is that really? Wima, ERO etc makes excellent products, at least if we talk industrial quality. They will never let you down.
But the coolest cap around is Electronc Concepts. I wonder why audiophiles haven't discovered it yet. Extremely heavy duty. Best in the world?
http://www.eci-capacitors.com/
http://www.diyaudio.com/forums/show...8221&highlight= |
|
|
| Nuuk |
| quote: | | But the coolest cap around is Electronc Concepts. I wonder why audiophiles haven't discovered it yet. |
Per, in that thread, you had not yet heard one! Seems strange to recommend a component for hi-fi if you have not actually heard it (even if you do live in the same city as Batman ;) ).
Have you had a chance to actually hear one yet, and if so, did it live up to expectations? |
|
|
| carlosfm |
I've always used "normal" polyesters on my GCs (from 2.2 to 4.7uf / 100v), and always with very good results.;)
Guys, you can start here and go up to something more exotic ONLY if you feel the need.
I don't.:) |
|
|
| Stever |
Have you checked out
http://www.partsconnexion.com/
Black Gate N 4.7, 50v - $US3.15
I built my GC with parts from above link and DigiKey. Both delivered within two days. |
|
|
| jean-paul |
| quote: | Originally posted by carlosfm
I've always used "normal" polyesters on my GCs (from 2.2 to 4.7uf / 100v), and always with very good results.;)
Guys, you can start here and go up to something more exotic ONLY if you feel the need.
I don't.:) |
What's wrong with *excellent* results using carefully selected quality parts ?
Why build an amp yourself, discuss inverting and non inverting topology and then use MKT when superior MKP is around ? The price can't be an argument as the price-difference in not worth mentioning.
Why not go for the best after the decision is made to build a GC ? The total price will still be very low so I don't see the point of using standard parts with only a handfull of components... |
|
|
| toque_101 |
| Why use sub-exotic parts? I'm a student and I don't necessarily have the $$ to buy really nice parts. When I can get a decent input cap from solen for $1.70 CAD vs getting a black gate cap for $3.50 USD (about $4.50 CAD) plus shipping, suddenly having only a decent input cap doesn't seem all that bad as long as it doesn't murder the sound. Keep in mind that I'm not totally skimping on parts, just not going for the top of the line stuff. This is also my first GC so I had to buy a transformer, case, etc. My budget is a little stretched :) Thanks for the help guys, I really do appreciate it! |
|
|
| GregGC |
| quote: | Originally posted by toque_101
Why use sub-exotic parts? I'm a student and I don't necessarily have the $$ to buy really nice parts. When I can get a decent input cap from solen for $1.70 CAD vs getting a black gate cap for $3.50 USD (about $4.50 CAD) plus shipping, suddenly having only a decent input cap doesn't seem all that bad as long as it doesn't murder the sound. Keep in mind that I'm not totally skimping on parts, just not going for the top of the line stuff. This is also my first GC so I had to buy a transformer, case, etc. My budget is a little stretched :) Thanks for the help guys, I really do appreciate it! |
I'd say loose the input cap totally (if no DC on the source output). Keep it as simple as possible. |
|
|
| toque_101 |
| Well, the source for the time being will be my pc soundcard so there's bound to be a bit of DC there ;) |
|
|
| analog_sa |
| There are bound to be exceptions but standard polyester types like WIMA MKS are plain ugly sounding and in fact worse than many elcos. Even the not-so-spectacular metalised WIMA MKP are incredibly better. Are you guys really counting the cents? |
|
|
| Nuuk |
The difference in cost between using the cheapest parts and using something to produce a very good Gainclone is about 5UKP (even when including the input caps).
I'm long-term unwaged due to ill-health and have to watch every single penny but I'm not stupid enough to spend a week building an amp to have it sound average when for an extra five pounds, it can sound wonderful!
We have a saying in the UK about 'spoiling the ship for a ha'pence of tar'. |
|
|
| GregGC |
| quote: | Originally posted by toque_101
Well, the source for the time being will be my pc soundcard so there's bound to be a bit of DC there ;) |
I don't want to argue with you, but most likely they have a cap at the output. So I wouldn't be so sure. In fact 99.99% I'd say there is no DC there. Prove me wrong (measure the output when no AC signal is present). |
|
|
| 2Bak |
I DC-measured from line out of my Thule pre-amp today and got:
Left ch: 0,1 mv
Right ch: 2,3 mv
Should I worry about these 2,3 mv if I leave out the 2,2 uf input cap in my GC power-amp?
What would critical values be around..?
/Jan |
|
|
| analog_sa |
| quote: | | Should I worry about these 2,3 mv if I leave out the 2,2 uf input cap in my GC power-amp? |
Yes, you should worry. The 2.3mV assuming gain x30 will result 60-70mV at output which is not disastrous but unless you've seen the circuit of your pre there is no way to know if this offset is temperature related and whether at power up/down the offset is not much higher. |
|
|
| GregGC |
| quote: | Originally posted by 2Bak
I DC-measured from line out of my Thule pre-amp today and got:
Left ch: 0,1 mv
Right ch: 2,3 mv
Should I worry about these 2,3 mv if I leave out the 2,2 uf input cap in my GC power-amp?
What would critical values be around..?
/Jan |
analog_sa/Jan is right. Put the cap in place so you sleep well. |
|
|
| officeboy |
I also did this, but other posts in this thread have me thinking
Aren’t BG N's electrolytic type capacitors? Are they suitable for input filtering? |
|
|
| Lars Clausen |
10 years ago i would agree that a PP cap is the best for this purpose. However today there are better alternatives.
I would point to the Black Gate NX Hi-Q, for instance 47uF 6.3 V Bipolar. It is small, it is fast, and does not color the sound at all.
I think prices are around 2-3$.
A more experimental approach is using a normal 10c electrolytic cap of 4.7uF, and bypass it (through resistors) with small MICA or Polystyrene caps. 100pF - 1000pF. This can give very good results, and you have some wheels to fine tune your preferred sound. R1 and R2 should be something like 47 - 100 Ohms |
|
|
| toque_101 |
alright, you guys have convinced me to do some more research and find out exactly how its gonna cost me to go with some good components.
GregGC:
I would like to measure the DC offset at the output of my soundcard, but my mulitmeter is a POS (will only measure down to 2.5Vdc) :) I have a new one on the way, but I'm going to have to wait for a bit. Anyone have a soundblaster live and can measure DC offset on it for me :D |
|
|
| officeboy |
| The older SB live in the computer i'm at now, at max volume is from 3.9mv DC with a 50hz sine. (it's about 1V AC) |
|
|
| GregGC |
| quote: | Originally posted by officeboy
The older SB live in the computer i'm at now, at max volume is from 3.9mv DC with a 50hz sine. (it's about 1V AC) |
Some times the noise on the output can give you small DC readings which is most likely the case. Just touch the leads of the DMM when on the lowest DC setting and you'll read some values. Connect a 1k-10k res in paralel to the SB out and see what you get. |
|
|
| officeboy |
| quote: | Originally posted by GregGC
Some times the noise on the output can give you small DC readings which is most likely the case. Just touch the leads of the DMM when on the lowest DC setting and you'll read some values. Connect a 1k-10k res in paralel to the SB out and see what you get. |
With my DMM, if the leads are floating, I get some readings, but once they touch I get a dead flat reading. That 3.9mv was also dead stable over 10 or so seconds. (after I remembered to take sample with sine wave cus the music was having it jump all over the place)
When I get back home I’ll connect something across and see what it reads. |
|
|
| GregGC |
| quote: | Originally posted by officeboy
With my DMM, if the leads are floating, I get some readings, but once they touch I get a dead flat reading. That 3.9mv was also dead stable over 10 or so seconds. (after I remembered to take sample with sine wave cus the music was having it jump all over the place)
When I get back home I’ll connect something across and see what it reads. |
Do you say you had sin wave signal comming out of the PC SC? |
|
|
| Christer |
Just a comment, don't expect the DC offset of one particular
soundblaster card to say much about the offset of another one.
Unless there is a tech. spec. giving max min levels, I would say
we know nothing even if we have meausements from other cards
of the same model.
BTW, my Soundblaster is definitely AC coupled, either at the output
or the input or both. I tried outputting a DC signal running loopback
and got zero volts back. |
|
|
| officeboy |
| Yeah, a 50hz sine wave file I still had in my winamp playlist from breaking in some drivers. I've found that my dmm averages the reading. |
|
|
| Christer |
| quote: | Originally posted by officeboy
Yeah, a 50hz sine wave file I still had in my winamp playlist from breaking in some drivers. I've found that my dmm averages the reading. |
Are you sure it averages to zero Volts exaclty when there actually
is no DC offset? A few mV could most likely be a measurement
error IMHO. Try the loopback method with a pure DC signal as I
sugggested and see what your soundcard captures. It won't
tell where there are caps, though. |
|
|
| officeboy |
| quote: | Originally posted by Christer
A few mV could most likely be a measurement
error IMHO.
Try the loopback method with a pure DC signal as I
sugggested and see what your soundcard captures. |
The Dmm I use reads accurately down to 0.1mv so I doubt a measurement error more then a few % +/- (on the meters part, but the thing is 3 times smarter then me).
How would I do a DC loopback test? Connect an AA battery to the input line? |
|
|
| Christer |
| quote: | Originally posted by officeboy
The Dmm I use reads accurately down to 0.1mv so I doubt a measurement error more then a few % +/- (on the meters part, but the thing is 3 times smarter then me).
How would I do a DC loopback test? Connect an AA battery to the input line? |
That would only test the input, which is interesting too, but it is
not loopback. Loopback means connecting the output directly
to the input. I then output a DC signal and checked what I got
back. The result was zero volts. You could do the battery test
(with a voltage divider preferrably) to check if the input is AC-coupled.
Then connect the DMM to the output and output a DC signal to
check if also the output is AC coupled. |
|
|
| officeboy |
So your just using the soundcards input as resistance for the line out?
What are you using for an output signal?
Should be easy enough to test on my Speaker Workshop computer. |
|
|
| Christer |
| quote: | Originally posted by officeboy
So your just using the soundcards input as resistance for the line out?
What are you using for an output signal?
Should be easy enough to test on my Speaker Workshop computer. |
No, I send a wave stream containing just a DC level to the soundcard
output and check what I get back from its input, in my case it there
was no DC level in the resulting wave stream. Ergo, there must be
at least one blocking cap somewhere along the line. |
|
|
| officeboy |
So you bombard your soundcard with a dc signal to see if it will amplify that signal? seems... er… very thorough. :dodgy:
I don’t think I’ll be doing that. An input filter could be blocking all the DC, and then your soundcard would see nothing to amplify, if I had no signal to amp I bet I would have 0mv DC as well.
Better to have a known signal and test with that. |
|
|
| Christer |
| quote: | Originally posted by officeboy
So you bombard your soundcard with a dc signal to see if it will amplify that signal? seems... er… very thorough. :dodgy:
I don’t think I’ll be doing that. An input filter could be blocking all the DC, and then your soundcard would see nothing to amplify, if I had no signal to amp I bet I would have 0mv DC as well.
Better to have a known signal and test with that. |
Nah, it wouldn't amplify, just reproduce. The point was to test if
it was DC-coupled or not. Since it was obviously AC coupled, there
shouldn't be any DC offset. OK, I neved did check if both input and
output were AC coupled. My purpose was slightly different, though.
If it had been DC coupled it could have been used for certain
interesting measurements. |
|
|
|
