Hi Paul, thanks for the input.Paul Barker wrote: ↑Mon Jul 04, 2022 11:20 am Steve
...I’m saying the enemy you perceive may not be the biggest issue and you’re feedback might make it worse......
The thing is, the feedback hasn't made anything worse, far from it. The proper application of the feedback had a massively positive impact on the quality of the sound. I know that I might have had a tendency in the past to either build and hope, or copy other people (Alex Kitic comes to mind on that one) but on this occasion I have been through the whole initial process on my own, with old books, a fine tooth comb and a terrible trial by mathematics to put it mildly. Frankly with the maths, I was on the point of giving up, but I gave myself a bollocking and determined that this time I wasn't going to bail out and was going to put in the effort, and sweat required to learn something useful.
Once I had the basic process down and had built something, I tested it with a few records and was frankly gobsmacked at how good the sound was. This was before any scoping of waveforms, poles, zeroes, or HF phase shift compensation calculations had even been thought about. That was another load of maths and I needed a tiny steps strategy to avoid getting overwhelmed.
I was certainly gratified that the maths, theory, whatever you want to call it I had learned thus far, had worked the way Crowhurst et al had said it would and because the sound was so good, I was satisfied that I was on the right lines. You've no idea how that felt after so many years of frustration. Let's just say there was a great deal of fist pumping, hooting and hollering when the first tune struck up from the speakers. It's a good job Melanie was out and the neighbours were on holiday. If she'd seen and heard me, I'm sure she would have been on the phone to the men in white coats toot sweet.
As you know, the HF phase shifts and where they occur are unique to every make of output transformer and although with a bit of estimation from the published specs and some rough working, it is possible to get in the right ballpark, the fine tuning becomes an iterative process. The mess of harmonics on top of the first 10KHz square wave trace without any phase compensation looked pretty damned ugly, and confirmed the need to do something about them: I managed to improve the traces with a worked out then tweaked up cap across the feedback resistor: After that, it was a matter of dealing with a bit of slew rate limiting as best I could, then fitting another worked out and tweaked up compensation network to the input stage anode to finish up with this: As I edged towards the best I could get out of the OPT, the sound just got better and better, to the point where now I have something that sounds great, but more importantly, I know why it sounds so good because I've used the triple combination of measuring, interpreting (with help) the test results and final subjective auditioning to get what I want.
It's the first time I've followed the process and now appreciate the reasoning behind it. The maths, the measurements and the advice helped me to define a clear set of goals, then make them happen.