First off, great video! As far as the caps, any X/Y rated caps are made to fail by internally vaporizing the metallic surface on the internal foil ( Dave @ EEVblog took one apart and showed the damage process), and will also drop in capacitance as they do that. Using them in AC power / dropper circuits like that they commonly fail over time, but maybe using 2 larger value caps in series would delay that by lowering the stress on them. I would also consider replacing those multiple board mounted resistors with a chassis mounted one with a higher rating, or, at least standing off those resistors higher on the board and maybe putting a closely rated slow-blow fuse inline with them to use as a fail safe in case any thing like that happens again. Using a ICL component instead of resistors, along with a slow blow fuse might also work nicely, as the ICL would take stress off of the fuse until the relay bypassed it, and a closely rated fuse would also open and keep the ICL from cooking things.

@bansci

4 ай бұрын

Thanks for the thoughts, all great points!

X and Y capacitors are rated for AC RMS mains voltage, so, if you have 230 v AC, then 275 v capacitor is exactly right for this voltage. Some X/Y capacitors are even marked as 250 v, it means, that it is specifically designed to work in 220 - 230 voltage mains circuit (even if peak voltage at 230 v mains is around 320 v).

Awesome stuff ! just watched your initial unboxing and thoughts on the 8 1/2 digit fluke and now this one and I have subbed. Audio and metrology are 2 things I enjoy but I am very much the amateur (got to start somewhere !!) and enjoy your delivery of what you do. I am at the moment fascinated by Sinewaves and making low distortion sources and in particular Bulb Stabilised Wien Bridge Oscillators. I have a crappy sound card and have been using ARTA and REW software to find my feet, along side my Siglent's FFT which is only realy an indicator that things are improving ! anyhow great work I look forward to more stuff like this....cheers.

@bansci

4 ай бұрын

I think low-distortion-nuttery is just as addictive as volt-nuttery, or time-nuttery, except it's easy more expensive as you start chasing the hertz, and gigahertz... Save yourself!

@andymouse

4 ай бұрын

Lol ! but but but the sinewaves ? :)@@bansci

Impressive. Thanks.

very interesting, thanks

@bansci

2 жыл бұрын

Glad you enjoyed it, thanks!

The JFETs 2SJ74BL and 2SK170 notoriously unattainable That's the only reason I didn't build this awesome amp

@bansci

Жыл бұрын

diyaudiostore.com/collections/frontpage/products/matched-jfets?variant=39335247020105 Now you have no excuse :-P

@davidaz4933

Жыл бұрын

@@bansci Thanks I am registered on the DIYAUDIO website many years I posted a lot of posts there about a DAC I built At the time they sold in the store they did not make deliveries to Israel I see they fixed it I checked now including shipping $47 I will later order a set from them I am currently more focused on the project of improvements to my spectrum analyzer R&S Looking for a good source for a 10 MHz rubidium reference I regret that I did not purchase it in the time that was on stock from a supplier tech mart in Israel for less than 50 dollars not working it very easy to fix them

@bansci

Жыл бұрын

@@davidaz4933 Ah you're in Israel? I always see excellent test equipment for sale from there on eBay, but the taxes and shipping are too much to the UK. Regarding frequency references, I really like BG7TBL's GPSDO, it has a second hand quality ovenized crystal in it, mine was a Trimble. He also does a frequency analyser too which seems well respected (called the FA2). The GPSDO needs to be left on, but I found it as good as my rubidium as far as I could test.

Very nice amp. What was the reason for the soft start over say two CL60's on the mains and transformer primaries? Must have more capacitance than typical.

@bansci

Жыл бұрын

Great question, I would say efficiency, but the amp is not very efficient anyway. I originally used it just because it was part of diyaudio's universal power supply design but given the trouble I've had something like a NTC solution is looking very attractive!

Nice amplifier, not sure it's really worth the inefficiency of running class a though. Class ab seems to give low distortion on most amplifiers I've tested. I use a hp 8903 audio analyzer for measuring thd. Can't remember what it's lowest resolution is but good enough for me

@bansci

2 жыл бұрын

Hearing is believing! I used to think the same but we did blind tests with 5 amps (arkham, cambridge, marantz, pass F5, ACA) and a selection of speakers and sources (LPs, DACs, CD players). You could hear the F5 was head an shoulders above the rest, especially with full range speakers. Not so much difference to my ear in the sources (aside from additional noise with the LP). All of this is blasphemy to one side or another in the "holy wars" of audiophilery. Is class A justified in my setup? Absolutely not, my room is narrow and the Mission bookshelf speakers can't match the F5. The amp is massive overkill. But this hobby is a trajectory, never an endpoint. BTW the HP 8903 is a beautiful piece of kit. So expensive on the second hand market. 0.01% distortion, so higher than the soundcard I show, but capable of measuring the F5's performance and with much more capability and with range switching. Would love to own one, just can't get it for a reasonable price!

@ptronix

2 жыл бұрын

@@bansci hi again, just checked the manual on the 8903, says its range is 0.001 - 100% , but I agree a decent sound card is probably the way to go. Think I paid less than £400 for my 8903, just lucky to find it on ebay, it had the connectors on the back like a lot do but is easy enough to put them on the front as the cables are quite long

@bansci

2 жыл бұрын

@@ptronix that's a great price! I've tried to pick up one a few times now but nothing less than £600, one even went for £1k! The 0.001% is the display range on the HP 8903, the analyser has a residual distortions and noise spec of 0.01% in the most sensitive range. This is actually amazing given its THD+N. The soundcard is using a Fourier transform trick to isolate the THD from the noise to get better specs (not sure about the 8903). I also read that most of the 8903's easily beat their specifications, so I'd much prefer one over my soundcard. But the sound cards are very capable, especially with REW.