Hi,Dxvideo said:
I have used them in my amplifier power supply (500VA, 88,000uF) so they will be more than adequate for a gainclone I think. I was recommended the MUR but they were too expensive so I used the BYW80-200 from Ebay. I got the idea from this page: http://www.hificollective.co.uk/components/diodes.html
I used parallel pairs at first but found that to be completely unnecessary. I used some small clip on heatsinks. 35nS is very fast. I cannot honestly comment on the 'sound' of these diodes because I also increased and updated the capacitance at the same time. But my amplifier now sounds better than ever!
Picture taken before removing parallel diodes:
Attachments
Dxvideo said:
I'm not exactly sure if they would be necessary, but, why take the risk when clip-on TO-220 heatsinks are so cheap?
One of my worries was the large currents that flow when first switching the amp on and charging the capacitors. I use 88,000uF so I use large diodes with heatsinks just to be very safe.
I don't like to take risks with power supplies but I may be trying too hard!!
Regards,
Martin.
Why not?
I can use the small heatsinks for 78XX series. But my snubbers are too small if compares to yours. I use only 4x4,700µ on each rail! And 2,200µ's on amp board..
Does that cause any hum?
And another question; should i put parallel hundreds nF caps to diodes? Because i saw this application on many GC psus.. In fact i dont know the reason..
Thanks.
I can use the small heatsinks for 78XX series. But my snubbers are too small if compares to yours. I use only 4x4,700µ on each rail! And 2,200µ's on amp board..
Does that cause any hum?
And another question; should i put parallel hundreds nF caps to diodes? Because i saw this application on many GC psus.. In fact i dont know the reason..
Thanks.
Dxvideo said:
I ask again: Why not use small and very cheap heatsinks?
The size of your capacitors are a personal choice. Some people believe less is better with gainclones. I think that, for amplifiers in general, more is better for low frequencies. I do not own a gainclone though so what do I know?
I believe the 100nF caps (on the diodes) help to suppress any high frequency switching noise. You can probably see from my picture that I now have spaces for these since removing the parallelled diodes. I will definitely be trying it very soon!! I'll let you know if I hear a difference.
You should try it too!! It is cheap, quite safe, and can be removed easily. This is what DIY is all about!!
Good luck,
Martin.
Sonusthree said:
There is no gain at all using small heatsinks to counter turn-on surge current. That surge is so fast the diode die has no chance to heat up the 'sink tab sufficiently and the duration is so short it won't matter. You would blow the diode from overcurrent long before a few ms surge would depend on having a sink or not.
As to "why not", it is simple, There is no end to the things we could add to an amp on superstition alone, but it would be without any useful purpose in some cases. To add something without purpose is madness, but that does not mean there is never a purpose. You are correct that it could need a heatsink, it is not unusual to need one.
The maximum average current through the diode, multiplied by the diode forward voltage drop (see the diode's datasheet for (forward_voltage:current rating) determines the heat generation. For a practical example, suppose MUR860 and 3A average current. The forward voltage in that scenario is roughly (depends on temp too, let's factor for about 60C or so) 0.75V.
0.75V * 3A = 2.25W
2.25W is too much for the TO220 packaged MUR860, it will need a heatsink in this scenario. Depending on ambient temp, many consider a TO220 package good for about 1W, but at least a diode tends to have a little higher max temp than more elaborate ICs.
However, if someone is using a pair of bridges, one per channel as often seen here, 3A average current through each diode has to be an amp with quite a load, I doubt most people who can still hear fairly well listen at this level. Many people are not building the rest of their amp for that kind of continuous output and are even building with linear regulators that aren't sunk well for 3X+ heat production at 3A continuous current, that alone signals a lower current amp design. BUT, overkill is sometimes what it's all about. Building a design with giant margins is not bad when it doesn't have a heavy penalty in cost, time, space, or performance.
It would be a bit mismatched to start 'sinking diodes for more current than the rest of the amp can safely handle continuously, rather than upgrading everything - but the builder had in theory already chosen not to build it that robustly for whatever reason (perhaps they don't need that much power, a chipamp that can output 60W can be build to be a 20W amp or whatever, in which case just as the diodes you 'sink are a saftey measure, so would using a higher current chipamp in a design than what is demanded of it).
Some people will choose a TO220 diode specifically because they don't want to use heatsinks. They are contrasting a smaller package axial diode or diode pack with no integral heatsink tab, versus the TO220 that has that tab for the addt'l benefit, better dissipation. There are lower-end commercial amps that only used 3A axial diodes and were rated for 200W output, may run for years before those diodes fail. Even longer if they'd had a TO220 w/o heatsink, the diode may never fail.
Yes, if you have room for a heatsink you might as well add one, but frankly I don't like the clip-ons at all, might as well just use a board mounted or a piece of thick Al sheeting and screw it on. The clip-ons tend to be thin, and a poor interface. If we concede a heatsink is useful, might as well make it the best heatsink that will fit in the space available.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.