Erik Mandaville Posted August 4, 2003 Share Posted August 4, 2003 I'm home sick today -- actually very painful hip after a bike crash yesterday morning...so I'm going to take the time to experiment with cathode resistor bypass caps on input and output stages. I built mine without these, since I was using the W.E. output connection with a 57uf capacitor. Jeff does not use a bypass capacitor on the input stage, and Leo took his out of the Moondogs. I also tried my Moondogs without; liked it more at first because of greater sense of 'air'; but then put them back in again. Chris and Ed already have the bypass caps in place -- at least on the 2A3 cathode resistor. The input stage is really a toss-up and subjective choice sort-of-thing. Reasons for experimentation: The Horus is a very quick and agile amplifier, and in my opinion tends toward the brighter side in terms of presentation. For me, this is a great advantage and the main thing I like about the design. Without the bypass cap in place, there is perhaps just a very slight lack of body. This may have been just a fleeting impression, but comes from listening to the amps at length over the weekend. It may be a capacitor value issue, and this could very well be one way to obtain some small control over the voicing of the amplifiers to one's preference. I'm going to install 100uf across the 2A3 cathode Resistor today to see what happens, as well as approximately the same value on the input stage. These are not going to be 'audiophile grade capcitors' (yuk!), but just some inexpensive electrolytics to listen to the alteration in sound. If things are improved, then maybe I'll get some Cerafines, or something like that. BTW: if you are familiar with the Bottlehead site, there is a recent set of posts there (with outstanding illustrations)of a DIY 2A3 amp using chassis from IAG. This guy's work, both inside and out, is superb. www.bottlehead.com (think that's the URL) go to the forum link, and scroll down just a little bit. Quote Link to comment Share on other sites More sharing options...
leok Posted August 4, 2003 Share Posted August 4, 2003 Erik, By the time I finished your note I didn't know which circuit(s) you are refering to .. maybe both. An interesting note: Over the past few days I've been studying, with Jeff's help involving various output configurations, where voltages and currents are going in my ultrapath parafeed design. The conclusion I came to is that the parafeed inductor is a constant current source for the two parallel paths: 2A3 / output-xfmr-ultrapath-cap. These paths recombine at the 2A3 cathode cap/resistor. Theoretically, the 2A3 cathode cap does nothing except handle 2A3 grid ac currents, otherwise (as is in the new Welborne amp design) the 2A3 cathode cap isn't needed at all. I believe a standard output transformer is theoretically the same, combining the parafeed choke and xfmr into a single unit. I told you I was wondering what that cap was really doing .. well, not much. Except that any voltage fluctuation due to grid currents and circuit imperfections will modulate the signal as seen by the grid relative to the 2A3 driver. So, in the end, I do think a very high quality, maybe small, 2A3 cathode cap is in order. All this assumes Ultrapath cap. Otherwise all output currents flow through the 2A3 cathode cap and the circuit is no longer constant current. Interesting huh? leok Quote Link to comment Share on other sites More sharing options...
Erik Mandaville Posted August 4, 2003 Author Share Posted August 4, 2003 I'm sorry for the confusion Leo. I haven't slept since night before last, and so maybe that's part of the reason for lack of clarity. I understand what you are saying --- thanks for taking the time to reply. Your assumption about that output capacitor connection is right, and I know that Jack Eliano's original DRD design doesn't involve bypassing of the cathode resistor. I know one other person other than yourself who has done this modification to his amps (also Moondogs), and said he preferred the sound (for whatever reason) with a combination of byapassing AND a slightly lower value capacitor between the bottom of the OPT primary and cathode. Jeff told me he thought it would be interesting to try it both ways, although I believe the quality of bypass capacitor becomes unimportant when using the Western Electric (sorry I prefer this to 'ultrapath')output connection. I think in AF circuits, not using a bypass capacitor can improve or reduce distortion, but sometimes at the cost of gain and/or sensitivity. (I remember reading about this in my RCA manual). If I remember correctly, Jeff tried the W.E. connection in his 300B amp (now for sale -- if this is the same amplifier), but did not specify if he removed the bypass capacitor, since it shouldn't be needed. But the output side of the circuit is largely taken care of, so what I think I will do is focus on the input stage. It's nothing to put a little capacitor in there to see what happens. As I said, I tried taking C5 out of the Moondogs, but put it back in again. My Moondogs are actually completely stock, now. CCS is/are interesting. I built a little preamplifier a few years ago that uses a single sub-mini tube (6021)for both channels. It usus small transistors (MOSFETS)to achieve active plate loading by approximating a constant current source. In this case, constant DC is pulled from the cathode, where the grid floats in DC potential, is able to bias itself, which in turn permits the grounding of the cathode without a cathode resistor and bypass capacitor. It is an extremely transparent preamp, but just offers much more gain than I need. I could use it when I had installed a volume control on the Moondogs, because the volume pot offered some control over the incoming signal. The possibilities with all of this seem endless... Quote Link to comment Share on other sites More sharing options...
leok Posted August 4, 2003 Share Posted August 4, 2003 Erik, The cc anode loading approach is new to me. Evidently not so new to you or Jeff. Interesting stuff. leok Quote Link to comment Share on other sites More sharing options...
Erik Mandaville Posted August 4, 2003 Author Share Posted August 4, 2003 I think just a little more clarification is in order for others who may read this: The term 'Ultrapath' is a term which was first used by Jack Eliano, of Electra-Print (company that makes audio transformers, chokes, etc.)Jack Eliano is the guy who designed the DRD amplifiers now offered by Welborne Labs. This so-called 'Ultrapath' connection is not new, having first been used by Western Electric years and years ago. It wasn't even called 'Ultrapath' then. It is just a really smart way of changing the position and type of capacitor that normally is connected across (bypassing)the cathode bias resistor -- in the 2A3, this is a high wattage power resistor. In a conventional output connection that DOES use a bypass capacitor across the cathode bias resistor, one end of the output transformer primary is connected to the plate of the output tube, and the other leg of the primary is connected (or returned) to the power supply. Jack Eliano's Ultrapath connection, the approach first used by W.E., returns the signal from the transformer to cathode instead of the power supply to cathode, which theoretically should offer better detail, less noise, etc. One of the more important advantages to this technique is that it effectively removes all electrolytic capacitors from the signal side of the circuit. That's why when cathode resistor bypass capacitors ARE used (as in a more tradtional SE output connection), it's good to use very high quality electrolytics. The Moondog, however, uses a lower value poly tubular capacitor in that position. I wanted to build the DRD amps a few years ago, but couldn't afford the transformers for them. It suddenly dawned on me that the Moondog would be a good candidate to try that 'ULTRAPATH' connection -- and so I did it. The thing was that, despite the fact that this can be done on a single-ended amplifier, there were other important aspect on the DRD design that make it quite a bit different from the Moondogs. Leok then tried the W.E. connection and really liked it. But he also had the insight (for which he deserves credit and thanks)to take this a step further -- and that further step being the employment of a parallel or shunt feed output in conjunction with the change in position of the cathode resistor capacitor. And this is the part that is of interest to me: Whether the W.E. (aka Ultrapath) connection behaves the same way with parallel feeding as it does with regular single-ended output transformer. In theory, it should. Parallel feeding basically frees the output transformer from having to 'pull two shifts'at the same time -- kind of like killing two birds with one stone (poor birds! My dad used to say to me, "You shoot it, you eat it!") I loved making different kinds of slingshots when I was a kid. So, that's some of the deal with the 'Ultrapath' connection. The other place a cathode bias resistor bypass capacitor (one with a lower voltage rating)is often used is on the input of the circuit. On the Horus, this is the 5687 tube -- a little guy compared to the big 2A3. The Horus does not use a capacitor in that position (it is not mandatory), nor does Leo on his Moondogs. Some have reported that they prefer the sound with the cap, others say it makes music sound a little less lively. What I'm hoping for, which among the main reasons the capacitor is used, is just a little more gain, which may translate into just a little more weight or heft to the presentation. I won't accept an alteration in the current high speed ability of this phenomenal amplifier, though. If the capacitor effects that characteristic in any way, it's coming out! I far prefer an airy and open sound to one with more bass -- I don't listen to music at really loud volume, anyway. If Segovia happened to be sitting in my listening room playing his guitar, I would want the volume level of playback from a CD or record to be similar. So. The End Quote Link to comment Share on other sites More sharing options...
Painful Reality Posted August 4, 2003 Share Posted August 4, 2003 ---------------- On 8/4/2003 8:28:10 AM leok wrote: Erik, The cc anode loading approach is new to me. Evidently not so new to you or Jeff. Interesting stuff. leok ---------------- CC anode loading will be somewhat new to me to. Gary Dahl suggested I took a look at Gary Pimm's pentode/mosfet constant current source (PMCCSS... phew!) for the 2A3PP on the drawing board. There is a nice text explaining the potential advantages of a CCS anode loading on a tube at http://home.pacifier.com/~gpimm/Active_loads_and_signal_current_control.html. It should be pretty fun to experiment with this. Quote Link to comment Share on other sites More sharing options...
Erik Mandaville Posted August 4, 2003 Author Share Posted August 4, 2003 I also know very little about CCS and active plate loads beyond what I've mentioned. I've used this technique in a couple of preamps, but never in a power amplifier. In both cases, those were someone else's designs. I have yet to do that -- I think I know enough now to maybe try, though. I'm OK with ceramics and sculpture (guess that's why I'm an art teacher!)but math is just something that never came easily for me. I have found electronics to be a great way to improve my understanding of it though. I wanted to build amps for years, but used to get horrified by really quite simple algebraic equations. It's not such a nightmare for me now. Of course in school my teacher didn't help much "Eveyone else can understand this, why can't you!!!" (in front of the whole class). I was mortified! Quote Link to comment Share on other sites More sharing options...
leok Posted August 4, 2003 Share Posted August 4, 2003 Dean, That's quite a reference. It completely separates the bias and output concepts. I wonder how much improvement such a circuit makes over a good inductor. leok Quote Link to comment Share on other sites More sharing options...
Painful Reality Posted August 4, 2003 Share Posted August 4, 2003 ---------------- On 8/4/2003 10:43:40 AM leok wrote: Dean, That's quite a reference. It completely separates the bias and output concepts. I wonder how much improvement such a circuit makes over a good inductor. leok ---------------- Dean or Jeff??? I wonder too how much improvement can be obtained over a good inductor. I guess the differences will be at the extremes. I once found a nice web page that did allow you to verify the frequecy response of a tube under choke loading, sadly I didn't put in in my "favorites" and I can't find the page back. As far as cost is involved a choke loading will be cheaper and much easier to implement. For sure we may be in the "law of diminushing return" realm here! I guess Lynn Olson would be the one to respond to this! Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.