Coupling cap installation

[Audioraider]

I have grown thoroughly confused as many comments in this forum seem contradictory. I am installing new caps in my Mainline and the manufacturer specifically states to install them “ The proper way to connect the outside foil is to the low impedance side of the circuit, which, in the case of coupling caps, will normally be the plate of the previous stage.” does this mean the outer foil of the cap connects to the tube side?
I assume the signal travels from the 6c45 tube to the transformer? If that’s the case the tube side (20 and 30 on the mainline) would be the low impedance side?

[Paul Birkeland]

Yes, the plate is the tube side.

[Audioraider]

I am going to install ODAM caps after all the spectacular reviews. Chris at V-cap asked me the impedance of the output transformer of the mainline?
You guys are the best and my next project after this will be the Crackatowoa

[Paul Birkeland]

The primary impedance of the OT in the Mainline is 4K.  I believe you'll hear a bigger difference from a nicer cap more than going from 10uF to 15uF.

[Audioraider]

Gotcha, after talking to Chris at V-cap and using his coupling Calculator he suggest I try the 18 UF ODAM in there. Since there’s very little difference between the price of 10 UF and 18 UF I’m gonna give it a try. according to his calculator an amplifier with a 4000 impedance output transformer would begin to roll off around 40 Hz with a 10uf cap. Not sure if that’s accurate but that’s his opinion. According to his calculator a 18 UF cap would be down 3db be at 19hz. After reading all the reviews online that the ODAM is one of the most transparent, warm and liquid caps ever made I’m really looking forward to hearing it.

[Paul Birkeland]

This will all depend a little bit on the actual headphones you're using and the output transformer itself.  Very few headphones will present their load impedance at low frequencies where you get driver resonances, so the headphone will unload the output transformer a bit down there and provide some extra extension.  Still, you could definitely try 20uF and there won't be any harm in doing so. 

[Audioraider]

The headphones are 300ohm ZMF Vérité. All my other headphones are collecting dust, the Utopia make for a very pretty paperweight😁
The thing that concerns me with the 18uf cap is the top end. There is a lot more material in the signal path then a 10uf but they are on their way so I will report back.

[Deke609]

Wait a sec. Are the cap values that Chris V. reported correct? As I read this thread, the OP is concerned about the -3dB low end cut-off frequency.  I may be wrong about this, but based on the numbers that Chris V. reported to the OP, my guess is that Chris used the formula for a RC circuit, or something very similar to it, where -3dB freq. = 1/(2*pi*C*R). Subbing in 4000 for R and 0.000001 for C (10 uF) gets us 39.8 Hz.  Subbing in 18uF instead of 10uF gets us 22 Hz, and 20uF gets us about 20 Hz. 

I don't really understand this stuff, but I would be shocked if PJ designed an amp spec'd for a low-end roll-off of 40Hz.  That doesn't make sense to me.  I see that in the Mainline the 6C45P is C4S-loaded, so I guess PJ's parafeed formula for choke-loaded output tubes doesn't apply. But John Broskie has a formula for minimum parafeed cap size to hit a target cut-off: C  = 159155/(rp + Rload)/Frequency, where rp is the plate resistance of the tube and Rload is the primary impedance of the output transformer. ( I think the 159155 number has something to do with the time constant of the cap). See here: http://tubecad.com/2014/09/blog0308.htm

I don't know the rp of the 6C45P at the Mainline operating point, but came across a reference to the rp ranging from 1200 to 2500 ohms. So, using 1200 and shooting for a -3dB freq. of 19 Hz, we get C = 159155/(1200 + 4000)/19 = a measly [edit] 1.6 uF, so call it 2 uF.

Elsewhere PJ suggested that the 10 uF stock cap size was chosen by experimentation. See here: http://forum.bottlehead.com/index.php?topic=12140.msg110652#msg110652 

Again, I may be wrong about all of the above. But if the OP only gets one kick at the can with these parafeed caps b/c of cost, and is only going for 18uF b/c of concern about cut-off frequency, then I think it would be helpful if someone with way more understanding than me ballparked the cut-off frequency of the stock Mainline when loaded with 300 ohms. 

Just trying to save the OP from choosing a cap value for the wrong reason.  And hoping I haven't badly confused the issue.

cheers, Derek

[edited to correct my poor arithmetic and typos]

[Audioraider]

That is exactly the questions I’m asking. I had initially ordered the 10uf until I found out the output transformer was 4K impedance.  I still have both the 10 and 18 on their way but I will have to make a decision without trying them because once I solder them and I can’t return and I have to return one of the two sets. When I asked Chris he said that if it was his amp with that output transformer he would use the 18 UF. I’m currently running the 15 UF Audyn reference caps and I can’t say that I’ve heard much of a difference in the bass although overall it is a substantially better sounding capacitor than the Dayton so I’m not comparing apples to apples. My goal is not to deviate from the designers intention but when I saw the 10uf rolling off at 40hz it confused me.

[mcandmar]

I think your off by an order of magnitude, 10uf is 0.000010F.  See http://sim.okawa-denshi.jp/en/CRhikeisan.htm  With 10uf into 4k i get a -3db of 3.97hz which seems spot on to me.  Anything less than 10hz i would be happy with.

[Deke609]

I think your off by an order of magnitude, 10uf is 0.000010F.  See http://sim.okawa-denshi.jp/en/CRhikeisan.htm  With 10uf into 4k i get a -3db of 3.97hz which seems spot on to me.  Anything less than 10hz i would be happy with.

Yup. Many thanks Mcandmar. I messed up the decimal place. (maybe Chris V. did too?  -- just kidding)  So, using the RC cutoff formula, we get 4 Hz with 10 uF, and 2 Hz with 20 uF. And using Broskie's formula, we get an even lower estimate of the required parafeed cap value. Neither formula is anywhere near the much higher capacitance values that were recommended to the OP.  So I don't get where those recommended values are coming from.

cheers, Derek

[Audioraider]

Thank you both for the clarification, that makes much more sense. Based on that and what Paul said in the other thread I see no reason to move larger than 10 UF because of bass response. I’ve read that you want to keep a cap in the signal path as small as possible to keep the signal as clean as possible but would there be any benefits to go up in cap size from the 10uf?

[Paul Birkeland]

I’ve read that you want to keep a cap in the signal path as small as possible to keep the signal as clean as possible but would there be any benefits to go up in cap size from the 10uf?

Oddly enough that isn't exactly always true either.  Atomic Bob did some measurements with coupling caps in the Crack and found some distortion decreases with larger values of coupling caps. 

If you're using 300 ohm headphones on the Mainline, one fun experiment might be paralleling a 100 ohm 1W resistor across each output while using the 300 ohm headphones.