[Brinkman]

So I've been amassing parts for my Castanet build and in the process have stumbled upon some tools that have been valuable aids in the picking of component upgrades.

The first is to know your load. Grados are typically of a nominal 32 ohm impedance, Sennheisers are in the 600 ohm range, and my AKG 701s are 64 ohms impedance. Lots of us can go to Headroom, look up our headphones and bring up it's Impedance vs Frequency chart. =703]Here's mine, looking to be closer to the upper 50s to 60 ohms of impedance (the 64 ohm listing is nominal).

So, with that info I calculated the reflected load on the primary of the stock output transformer (Hammond 119DA) with it's spec'd winding ratio of 12:1.
[(12*12)(60)]=8640 ohms 8.5K is close enough.

The Hammond 119DA is perfect as an off the shelf transformer for a kit such as Jim selling, though Sowter in the UK offers a headphone output transformer at a much higher cost. Instead, I contacted Jack at at Electra-Print and specified he wind a pair of 8.5K // 60 ohm 1W Parafeed output transformers. With Parafeed transformers there is no gap in the core (unlike SE and PP). This increases the inductance of the primary which translates to increased low-end handling.

I am also having him wind the pair of 40mA plate chokes (reactors). These will likely be in the realm of 70H.

I also found a highly informative webpage devoted to figuring values for Parafeed circuits. It's called Effects of Parafeed Parts Values. There's a useful Excel-based calculator you can download that is really helpful.

Also Excel-based (and a bit simpler to use) is a calculator program I found in this thread at the Intact Audio forum. It accomplishes much of the above calculator but is less in-depth and exacting.

Lastly, there's an online calculator here. It helps you calculate the resonant frequency of the modified inductance of your plate choke (stated L *0.9) and the Parafeed cap. You can also plug for a low resonant frequency (8Hz) and the modified inductance to find the cap size.

I'll keep posting as I complete my build.

Ben

[Brinkman]

So it turns out the footprint of the Electra-Print plate chokes would be too large for the Castanet PCB. So the plate chokes are going to be stock. The Electra-Print Parafeed OPT, on the other hand, are a perfect fit.

I am also going to try to integrate a 50K Stereo Goldpoint stepped attenuator I had originally purchased for my abandoned Clarinet build. This will replace the stock volume pot. Note: I had to increase the volume pot punch-out diameter of the Front Panel FPD file slightly to accommodate the Goldpoint.

Another change is to upgrade the carbon film grid stopper resistors (R11 L&R) with Kiwame carbon composition resistors. All other fixed resistors are stock. When the Kiwames arrive today I can start to finish my build.

When I purchased the Kiwame resistors, I also decided to bite the bullet on a NOS 6H30DR. These Reflektor-built 6H30s are supposedly an upgrade from the current production tubes. At any rate, this substitution was cheaper than a pair of Mundorf silver/oil parafeed caps and I am willing to speculate that it is just as noticeable an improvement.

Finally, I plan on experimenting with a tweak called DrP (damped resonance Parafeed). More can be read here. In short, it involves paralleling a series CR network with the Parafeed cap. The C is typically a cap twice the value of the parafeed cap in series with a large-ish value resistor. Specific values can be figured from the aforementioned link. The goal is to damp the LF resonance for a smoother, more controlled LF frequency response.

It is likely I will experiment with the values of the parafeed and DrP caps while trying to fine-tune the bass by ear. I may keep or reject the DrP. Either way, when I am satisfied with the value of the Parafeed cap (if I change it at all), I will then upgrade it with some Mundorf silver/oils, possibly bypassed with some FT-3s.

About the only other change was to substitute the TRS jacks with XLR jacks. While the Castanet circuit is not balanced, AKG701s are bi-wired and mine are terminated with male XLR plugs to allow compatibility with a balanced drive. Using XLR jacks spares me the need for a TRS-to-XLR/XLR adapter. The Front Panel FPD file also had to be tweaked to accommodate these XLR jacks, though I botched it and will have to mount both XLR jacks at an angle.

As the construction finishes, pictures and impressions will follow.

Ben

[flocchini]

Ben:


While you are  experimenting you might want to try 6H6∏ tubes. I am currently using these in my Castanet.Price and sound can't be beat. I use Mundorf Silver/Oil as the output cap.  I also thought about Jack's transformers so will be interested in what you think.

Best

Bob

[Brinkman]

Flocchini,

Thanks for the  6H6∏ tip. Do you know anything about how swappable this tube (aka 6H6P) is with the 6H30P? I can't find anything.

I got a lot of assembly done today.

However, it turns out I need to get specially threaded screws for my Neutrik XLR jacks. Not only that, but I may have to have the whole front panel rebuilt to correct my XLR jack mounting errors as well as scooting the volume pot punch out down for the footprint of my stepped attenuator. I don't think it will fit under the PCB correctly. Maybe a weeks delay.

BTW, Jack was really affordable; the Parafeed OPTs were only slightly more expensive than a pair of 3.3uF Mundorf silver/oils. From what I've read elsewhere, his OPTs are world-class. Can't wait to hear them.

P.S. I read an old post on this subject (Parafeed OPTs) where you were informed that Jack will need the winding ratio to build your OPTs. From first-hand experience, I can tell you he doesn't care about this. He wants only your desired Primary and Secondary impedances. For this, I simply copied the 12:1 ratio of the spec'd output transformer and applied it to a specific load - my AGK701s. So I then gave him the 8500:60 impedances, which ends up being a slightly smaller ratio than 12:1.

[Brinkman]

Holy sh!t.

I'm on my third front panel.
I made the jacks on the second panel too far apart for my headphone's XLR cable split.

Like the saying: in for a pence, in for a pound.

As for other changes I'm making, I decided to upgrade the 470uF cathode bypass capacitors (C9 L&R) to Elna Silmic IIs. Even if the ESR is higher for the Silmics than the spec'd stock types, I figure the 1uF bypass would mitigate that pretty well. I've used the Silmics in the past and like them. I'm still considering what to upgrade the bypasses (C8 L&R) with. The Elnas aren't a perfect fit and were best mounted on the bottom side of the PCB, barely leaving enough room for the stepped attenuator.

The 3.3uF parafeed caps (C10 L&R) have been upgraded to the white film & foil Solens. I could not find much on the internet as far as feedback is concerned in regards to these caps, but they are affordable for film & foil so they're worth a shot. They also weigh a lot for their size and that is encouraging for some strange reason. Despite their 150VDC rating, Jim said they would be fine, and if I like them, I may also replace the cathode bypass cap bypasses (C8 L&R) with them. Or some 1uF mundorf silver/oils I have reserved for my Clarinet build.

For the DrP caps (see above post) I went with the affordable 6.8uF Dayton audio metallized polypro. John Broskie considers these sonically superior to the ordinary metallized poly Solens. The series DrP resistor is a 10K PRP.  The goal is to damp the resonant spike of parafeed circuits that occurs at very low frequencies without damping the low end altogether. I don't have much time to fuss with this (and don't own a scope), so it will be a simple accept-or-reject experiment. If I don't hear an audible improvement, I'll yank it out.

My assembly should be finished Wednesday, with my impressions within a couple days of that.

[poty]

I wonder why don't you create your own PCB for the project. It is not so difficult (nor expensive) and you can fit all your specific parts with their sizes.

[Brinkman]

Poty,

The PCB is fine. It is the front panel of the chassis that posed the problem. Initially, I had no way of telling where the PCB was oriented with respect to the front panel (how much clearance there was above and below the PCB).

For the record, below the PCB there is 55mm clearance before the bottom of the chassis. The PCB itself is 2mm thick. Above the PCB is just shy of 36mm clearance before one hits the ceiling of the chassis.

So it was the dimensions of the attenuator and XLR jacks that required my redesigning of the front panel file. I had to make sure, as mounted to the front panel, they all fell beneath the PCB.

The only thing too large for the PCB was the Electra-Print plate choke. It would have been of a higher build quality but would likely have been the same in terms of inductance so I opted for the ease of stock, especially with my moving-out-of-state deadline looming before me (CDs and headphones only, no  soldering of kits).

Ben