Welcome! Need support, you got it. Or share you ideas and experiences.
I have had some great success with mods to the Ella that I think are relatively rare! Based on a hunch, I suppose, I tried bypassing the SRPP preamp stage with my Foreplay preamp. Amazingly, it worked! However, not at first. The fact that the driver tubes on both channels dont have the same pin numbers in the circuit fooled me for a while. However, with that fixed, the sound is amazing! The Ella seems to like using the Foreplay as its preamp circuit. I can't really compare before and after, however, since I changed speakers in the meantime.
Now for my question. Since I have taken out the 6922 stage in the circuit, does that free up more power in the transformer so I can bias my KT88s deeper into class A? Speaking of which, I have never really heard a ballpark max figure for biasing for the Ella. I suppose there is a large grey area. How would I calculate that considering the x-former rating of .6A at 315V?
Are you taking the output of the foreplay to the phasesplitter via a coupling cap or direct coupled from the cathode follower? You want around 100V on the splitters' grids for things to be working right. I assume your global feedback is disconnected now, or are you applying it around the phasesplitter?
The PTx will take a lot more current without complaint. You can push the HT past 325ma, but then your B+ sags to ~350V (need more filtering too w/o nfb). Good for the class A triode thing. Heater current can also be upped a little while still staying within 10% of 6.3V.
I basically modded this based on instinct, so it was simple and brutal. I have the capacitor-coupled output of the Foreplay tied directly to the phasesplitter of the 6sn7s. It sounds to be working quite nicely, however, despite the fact that the Foreplay is supposed to only give about 20v-not 100- max on its output.
My feedback is permanently disconnected, and I am thinking about doing that for the UL mode as well.
One of the things I was curious to try was to direct-couple the foreplay output to the phasesplitter. However, I suspect that might raise issues with the ICs carrying DC (seems like it should be bad).
How does one go about upping heater current?
What I meant was that the splitter's grids are meant to be at +100VDC because they are originally direct coupled from the SRPP stage. Having them near ground would be okay except that the common cathode resistor is 15K (iirc). So you are running the splitter into cutoff because its not biased correctly.. not enough current. You could separate the cathode resistor into a large "tail" resistor and a small bias resistor and connect 1M grid resistors to the junction. Otherwise, to keep that 15K resitor like it is, you would need to connect it to a negative rail instead of gnd.
Other than that, I'm sure the foreplay has enough gain to drive the amp minus the front-end stage. And I probably wouldn't run DC out on ICs either (just make sure you know what you're connecting it to).
OK, since I don't have the parts to implement the cap-coupled stage correctly yet, I gave a try at DC-coupling the Foreplay to the Ella. When measured with a DVM with no load, both the SRPP stage of the Ella and the Foreplay stage had a little less than 70 VDC on the outputs. So, I figured, why not? And, it seems that the sound is a little fuller, with more bass. I suppose this is because the 6sn7s are running a lot closer to their optimum. Anything I should watch out for here? It seems like there are probably circuit optimization issues to worry about... Any good sources of information to learn about this? Thanks,
You can use your preamp that way (DC to Ella 2nd stage), but the way I look at it, the stock 3 stage Ella is prefferable to using an external preamp (with 70VDC ICs). The attenuator and srpp stage does everything the foreplay does and its built right in. If you need more sources you could add those to the ella without too much trouble.
Other things that may be interesting. I changed the 6922 stage to a 6cg7 (mini 9pin 6sn7 approximately), which I like better than either the 6922 or the 12au7. You can also hook this first stage up in several ways. Don't have to be limited to srpp, though its not bad.
then you can improve the splitter/driver stage by forcing better balance with a constant current source. An easy way is using depletion mode mosfets in cascode. I like triodes for outputs, but you may have luck with the UL mode and no global feedback. I think I mentioned this before, but careful when you disconnect the feedback. If you disconnect it from OPT 2ndary side you will also disconnect a RC zobel network (.22uf +24R) that may be beneficial to sound when running in UL mode. Just my guess, but maybe someone can confirm..
I have to agree that the OPT seems have some high frequency ringing. In triode, it is there but not in your face as is UL. The zobel seems to help, don't recommend speakers with a flat treble response.
I have considered the interstage trans., but do not want to put any more $$ into this amp. I did the P-P rewire not too long ago and am pleased with the sound. The resistors are on the sockets, the amp images better and there seems to be more detail. Also added another cap for a CLC power supply filter. Tried various changes to the SRPP stage as well.
I might opt for a 2A3/300B project if I can find a speaker I like to mate the amp with. I seem to have an affection for 87db speakers.
The Foreplay must sound pretty close to the SRPP stage, and the only real reason I wanted to try it is because I have the Anticipation CCS installed on that. The only other possible advantage is the larger chassis space to add stuff in. My ideal implementation would be to switch those CCS boards from the Foreplay to the phase splitter stage, but I seem to have lost the manual which details how to alter the CCS specs. Maybe its time for a post in the Bottlehead forum!
Did you have to adjust many things in the first stage to convert to 6cg7?
I think that I will be sticking with triode mode and no NFB for now, as my new Fostexes allow that.
This is exciting, my first foray into real circuit design. I think I understand what you are talking about, but I am going to do a little research to make sure and then try the circuit adjustments tommorrow. I suspected that it wasn't optimal, the sound is really sweet in the mid and highs but the bass seems lacking ( I first thought that was due to my new DIY Fostex's). I'll let you know the progress.
This post is made possible by the generous support of people like you and our sponsors: