tube fanatic Posted April 4, 2014 Author Share Posted April 4, 2014 JL, are you talking about a build thread for the SET I just posted the picture of, or the forthcoming one with 6BX7s? My biggest problem is lack of time to spend on the computer each day, but I'll certainly try to entertain any requests for info whenever possible. Mark, you may be sorry that you asked! I developed this as a design concept for extremely low distortion (by triode standards), and it's a spin-off of the low power SEP described in the amp article that I posted a while back. It utilizes a separate p/s (my preference whenever the choice is mine) which brings in very low ripple dc (in this case, about 0.3% iirc), which runs not only the B+ busses (the amp is dual mono with only the input filter cap, located in the p/s, being shared by the two channels), but also the filaments as well. The latter are a series string arrangement. In spite of the parallel output tubes (50L6s), power output is only 600 mw/channel at clipping. While this may seem like a ridiculously low amount of power, it's way more than I need or ever use. I've demonstrated this amp in homes featuring CWs, K-horns, and LSs, and the listeners have refused to believe how low the power output is. The voltage amps are 12SJ7s- way overkill since only a bit over 5V rms is needed for full output, but I'm overrun with them and decided to use a pair (in addition to which their distortion at such a low output voltage is extremely low- around 0.5% according to Sylvania's published data). The tubes are biased at only -7.5V, and the plate voltage is 110 -- same parameters as when set up for SEP operation. Plate current is about 106 ma. Calculated second harmonic distortion is only 1.56% at full output. This is confirmed by measuring the load line distance between the quiescent operating point and double the bias, and dividing that into the measurement between zero bias and the the quiescent operating point. The ratio is approximately 1.1:1 (with a ratio of 1 corresponding to zero distortion, and 1.22 corresponding to 5%). If the opt tap corresponding to the minimum speaker impedance is used (4 ohms in my case), the distortion goes even lower with the usual higher speaker impedances. In terms of sound, the amp is remarkable. Huge soundstage, extreme definition of vocals and individual instruments, silky smoothness, and near total silence (approaching of that of the amp being turned off) under no signal conditions. The bottom end, due to the monster opts, is very potent as expected. The load impedance is 1.25k. So there it is! I'm hoping that someone on the forum is constructing the SEP variant featured in the article I posted. That also is quite remarkable, and triode conversion of that unit is very easy. With a little modification of the bias and plate voltage, it's possible to get almost a half watt from that amp when strapped for SET operation in spite of only one output tube. I keep one of those here for my own enjoyment. I've attached a picture of the load line used for the parallel SET in case anyone wants to look it over. It is graphed for a single tube load of 2.5k which, of course, needs to be halved for 2 tubes in parallel. Maynard Quote Link to comment Share on other sites More sharing options...
JL Sargent Posted April 5, 2014 Share Posted April 5, 2014 I'm talking about the 6BX7 build since that will be the upcoming project. Open it up for others to learn something? I realize that some people don't care to share much but you seem to enjoy sharing your knowledge and your work looks great. I just think it would be fun to "join" the build at some level. Might be too time consuming for you and I can understand that if it's the case. Quote Link to comment Share on other sites More sharing options...
tube fanatic Posted April 6, 2014 Author Share Posted April 6, 2014 On 4/2/2014 at 1:11 AM, mike stehr said: I was kind of wondering if you had a schematic of the parallel SEP, using the 6SF7 driver and 6AQ5 for output. I have a pair of James 6123HS universal output transformers, with 2.5K, 3.5K, and 5K primaries. The are 20 watts, but the primary current rating is 90 milliamps I was reading that thread of the parallel SEP amplifier, and you mentioned you were running the output transformer right at the max rated primary current...but the Tubelab guy tested and determined the Hammond OPT gave a extra 10% above the rating. I'm not sure if I should run at the max rated primary current, or exceed it with these James OPTs. Of course, they don't list those sort of things in spec/measurement data. I'd like to breadboard the circuit for some reason. Mike At Mike Stehr's request, I've attached the schematic of the parallel SEP amp which I posted a while ago. Since it consists of 2 separate amps on one chassis, only one channel is shown. Of course, it can be built as 2 separate units if desired. Some general info: R8, S1, and C5 comprise the optional "ear bleed" filter which can be used to make very harsh, over processed, recordings much more listenable. It's nothing more than a variable RC low-pass filter with a switch which allows it to be totally bypassed. If you don't want to install it, simply connect the input jack to R9 instead. Resistor wattages and capacitor working voltages are intentionally higher than necessary. It's just part of my overkill philosophy and adds very little cost vs. lower rated components. The power supply section came about because I happened to have the specified power xfmrs on hand. Any supply which can furnish approx. 250V @ 100 ma or more FOR EACH CHANNEL is fine. Also, be sure that the needed filament current of at least 1.5A PER CHANNEL is available. R15 is employed in the filament circuit to prevent filament voltage from exceeding 6.3. It's common for modern power xfmrs to provide as much as 7+ volts on the filament winding- this is way too high, and will result in greatly reduced tube life. The fuse value shown is needed because of the very large initial surge current when the amp is turned on. You can try reducing it to whatever value won't blow with repeated on/off cycles. The fuse, switch, and power cord can be shared if desired (in my unit I preferred to fuse each primary separately). C2 is a suppression cap which will prevent any pops from being heard in the speakers when the amp is turned on or off. Regarding the output xfmrs, any type can be used as long as it presents a 2.5k impedance to the output tubes with the speaker load connected. It must be able to handle a 90 ma, or greater, bias current. In reviewing the schematic and parts values, I don't think there are any mistakes (!?), but please contact me if anything seems to be amiss. Maynard Schematic deleted as the specified power transformer could not handle the peak current required by the bridge recitfier with capacitor input filter. Quote Link to comment Share on other sites More sharing options...
erik2A3 Posted April 7, 2014 Share Posted April 7, 2014 Maynard: thanks for the description, as well as for sharing the schematic. Very curious to experiment a bit with grid leak bias. Nice job on the build, by the way! Very pure DIY construction! Erik Quote Link to comment Share on other sites More sharing options...
mike stehr Posted April 8, 2014 Share Posted April 8, 2014 Thanks a bunch for posting that schematic, Maynard. It may be some time before I can get around to bread-boarding the circuit, but I will definitely do it. Quote Link to comment Share on other sites More sharing options...
Deang Posted April 8, 2014 Share Posted April 8, 2014 Really impressive. I wish I knew what in the hell you're talking about. : ) Quote Link to comment Share on other sites More sharing options...
Thaddeus Smith Posted August 4, 2015 Share Posted August 4, 2015 moved this thread here for relevance and fresh eyes. I now own this amplifier. :D 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.