Hi Tibi, can you explain, at a beginners level, what effects/changes you would expect by doing this?
Thanks,
Scott
Thanks,
Scott
I actually used a flatpac on a juma inspired F5:
F5 with 2SK2013/2SJ313
Measured really well too (attached).
I think the ripple was up around 200kHz if I remember correctly.
F5 with 2SK2013/2SJ313
Measured really well too (attached).
I think the ripple was up around 200kHz if I remember correctly.
Attachments
Yeah that's totally insane pricing. Not sure what's up with that ebay seller. But as Pchw mentioned, they are possible to get for $60-70, even less than that sometimes. If fact if you search for flatpac right now on eBay you'll see many for that price, just not in dual 24V configuration. You kind of have to scout around from time to time. Of course they're never around exactly when you need them lol, but search around once a month or so they def turn up cheap.
Hi Anchan,
those flatpacs still look kind of pricey - I mean, people around here scream about intolerably expensive Lateral MOSFETs for about 5 bucks a piece, let alone $500 for PSU...
(today's ebay screenshot)
Hm... I'm not much of a shopping hunter but if I set my wife onto it I bet she'll get them for 50 bucks
I've had not much time for DIY the past years but I did make 2 amps with Vicor Flatpacs. Both with amp boards designed by me and home etched.
Also I was running an NS10 preamp with a 24v Flatpac - this wasn't the greatest. Noisy. The flatpacs work better with a heavy load I THINK.
First amp was an M2 with 2 x 400VA Vicors in a separate chassis. Flatpacs are in series with center connection as 0v. Less than 10,000 uf capacitance. This amp is gooood.
The second amp is an Aleph 5; flatpacs are internal. Some sort of oscillation here and took out the tweeters on my Klipsch. I ended up with the C***** titanium which I don't like as much. This has 4 x 48v flatpacs trimmed down to 34v. 2 per channel in series center connection as 0v. The issue here could easily be my board design but I didn't take the time to look!
You can find the flatpacs on ebay reasonably if you keep looking.
Also I was running an NS10 preamp with a 24v Flatpac - this wasn't the greatest. Noisy. The flatpacs work better with a heavy load I THINK.
First amp was an M2 with 2 x 400VA Vicors in a separate chassis. Flatpacs are in series with center connection as 0v. Less than 10,000 uf capacitance. This amp is gooood.
The second amp is an Aleph 5; flatpacs are internal. Some sort of oscillation here and took out the tweeters on my Klipsch. I ended up with the C***** titanium which I don't like as much
. This has 4 x 48v flatpacs trimmed down to 34v. 2 per channel in series center connection as 0v. The issue here could easily be my board design but I didn't take the time to look!You can find the flatpacs on ebay reasonably if you keep looking.
This is a good topic I think - the SMPS thingy.
I as well look forward to post PSU filtering ideas. I think it's important as the noise at these switching frequencies can be difficult to predict when you consider harmonics and IMD (and Passive IMD). Maybe all we really need is a zobel @ the PSU output?
I think the iron in the M2 helps.
Never had the time to take it further....
I as well look forward to post PSU filtering ideas. I think it's important as the noise at these switching frequencies can be difficult to predict when you consider harmonics and IMD (and Passive IMD). Maybe all we really need is a zobel @ the PSU output?
I think the iron in the M2 helps.
Never had the time to take it further....
I also think discussing SMPS is a good topic. I always get poo-pooed when I bring it up lol but I think there are big advantages. Size and weight, but also it really gets rid of the 60/120/180Hz (or 50/100/150) noise right away. So now, we only have to focus on attenuating 100K or higher.
The risk is mainly that they can and do fail, and I think more often than a regular linear PSU.
In the case of the F5 variation, I did have concerns about what happens if one rail fails. I had +/- 15V rails, and what I did was use a relay on each speaker, where I ran +15 to one end of the coil, and -15 to the other, with a resistor to knock the voltage down to 24V across the coil. This way, speakers are only connected if both rails come up. Pretty remedial speaker protection, but seems to work pretty well.
I think speaker protection can work great if we anticipate what the most likely failure might be, and target our protection to exactly that.
The risk is mainly that they can and do fail, and I think more often than a regular linear PSU.
In the case of the F5 variation, I did have concerns about what happens if one rail fails. I had +/- 15V rails, and what I did was use a relay on each speaker, where I ran +15 to one end of the coil, and -15 to the other, with a resistor to knock the voltage down to 24V across the coil. This way, speakers are only connected if both rails come up. Pretty remedial speaker protection, but seems to work pretty well.
I think speaker protection can work great if we anticipate what the most likely failure might be, and target our protection to exactly that.
Got the higher current four pole filter ready for testing. PCB silkscreen says "6A max" but that's with the third best inductors. With these first best inductors, the ones shown in the photos, it'll handle 9 amperes just fine. Total in-to-out resistance is about 0.04 ohms.
It's still a prototype and it could need revision(s). But it's an example of what is possible to do.
It's still a prototype and it could need revision(s). But it's an example of what is possible to do.
Nice, I hope I get to play with those boards in a Sony VFET amp
But, I am interested in seeing how thinga trend with SMPS in high-end DIY audio. I recently completed tomchr's Neurochrome HP-2 as a gift for a friend, which leverages a Meanwell smps, and I couldn't really tell a major difference between this and my HP-1 which has a built-in psu.
I'll be keeping an eye on these recent builds taking advantage of SMPS.
But, I am interested in seeing how thinga trend with SMPS in high-end DIY audio. I recently completed tomchr's Neurochrome HP-2 as a gift for a friend, which leverages a Meanwell smps, and I couldn't really tell a major difference between this and my HP-1 which has a built-in psu.
I'll be keeping an eye on these recent builds taking advantage of SMPS.
Nice simple Class A amp, Juma. I also tried a SMPS in series to get +/-24v rails typically used on 25w Class A amps. I had a “forest of noise” issues with my linear PSU and when I tried the SMPS, it was so much cleaner.
Here is the FFT of my M2 with linear supply:
And here is with two 24v 5A supplies designed for LED lighting:
Here is the FFT of my M2 with linear supply:
And here is with two 24v 5A supplies designed for LED lighting: