I've been experimenting with building my own bridge rectifiers from four IXYS 6A 600V diodes, to go into an audio amplifier.
My amp came with a 25A standard bridge rectifier, 'someone' told me that summing four 6A diodes would be fine when building a bridge rectifier to replace it.
Exactly what are the effects at high volumes if the 6A IXYS diodes can't handle the current? Do they overheat, or would they tend to 'clip' on heavy bass notes?
My amp came with a 25A standard bridge rectifier, 'someone' told me that summing four 6A diodes would be fine when building a bridge rectifier to replace it.
Exactly what are the effects at high volumes if the 6A IXYS diodes can't handle the current? Do they overheat, or would they tend to 'clip' on heavy bass notes?
I don´t really think that paralelling diodes in PSU rectifier is a good practice.
The diodes will probably different forward voltage drops, so one of them will carry the highest current, while the others a lot less.
Something will probably give up.
The diodes will probably different forward voltage drops, so one of them will carry the highest current, while the others a lot less.
Something will probably give up.
reiver said:
No... not a good idea at all! Use a standard recovery diode bridge similar to stock.
Using four 6A rectifiers of any variety to make a bridge results in a 12A capability, not 24A. They don't all conduct at the same time.
The use of high speed diodes in the rectifier bridge of standard 60 Hz power suppiles is part of the pseudo-science surrounding the latest generations of chip amps. They all seem to do it. The high charging currents of overkill electrolytics are murder on high speed rectifiers and tends to reduce reliability. If a rectifier bridge is generating too much RF for your liking, the proper soulition is a SNUBBER, not faster diodes.
The use of high speed diodes in the rectifier bridge of standard 60 Hz power suppiles is part of the pseudo-science surrounding the latest generations of chip amps. They all seem to do it. The high charging currents of overkill electrolytics are murder on high speed rectifiers and tends to reduce reliability. If a rectifier bridge is generating too much RF for your liking, the proper soulition is a SNUBBER, not faster diodes.
This is all pseudo-sciencie.
Some of the diodes will just become shorted if they are stressed enough past their ratings. There aren't any other effects.
Some of the diodes will just become shorted if they are stressed enough past their ratings. There aren't any other effects.
wg_ski said:
Would have thought it would be 6A in fact? Each diode can have no more than 6A current through it and having 2 in series as they effectively are in a bridge won't change this. In parallel it should go up to 12A but parallel diodes is bad practice
re:
That's kinda what I wanted to hear. I'm running 40,000mf of caps in the amps power supply, and four 6A diodes running as a bridge rectifier scares me a bit.
Vishay makes a quality "stock" 25A bridge that I've used before, I think I'm going to forgo the Schottky hype and switch back.
60hz isn't exactly the high speed switching that it's intended to help anyway...
-Bryan
That's kinda what I wanted to hear. I'm running 40,000mf of caps in the amps power supply, and four 6A diodes running as a bridge rectifier scares me a bit.
Vishay makes a quality "stock" 25A bridge that I've used before, I think I'm going to forgo the Schottky hype and switch back.
60hz isn't exactly the high speed switching that it's intended to help anyway...
-Bryan
getting only what you pay for
It's not parallel because the diodes don't conduct at the same time. When you buy a 25A bridge, the individual diodes are "12.5A Average forward rectified current. It is assumed that the bridge will be driven with AC, with one pair conducting on one half cycle and the other pair conducting on the other half cycle. That puts half the average load current in each leg. If you want 25A diodes, you gotta buy the 50A bridge.
Dr.EM said:
Would have thought it would be 6A in fact? Each diode can have no more than 6A current through it and having 2 in series as they effectively are in a bridge won't change this. In parallel it should go up to 12A but parallel diodes is bad practice
It's not parallel because the diodes don't conduct at the same time. When you buy a 25A bridge, the individual diodes are "12.5A Average forward rectified current. It is assumed that the bridge will be driven with AC, with one pair conducting on one half cycle and the other pair conducting on the other half cycle. That puts half the average load current in each leg. If you want 25A diodes, you gotta buy the 50A bridge.
Your average currenty is probably just 1 or 2 amps. Running a 100W amp with +/-45V supplies will, if you run it full power on a testbench, require something like 2-3 A depending on the amp. But that's under test conditions, under music conditions surely less than an amp. Do the math.
The reason why they use those 25A bridges is because the peak currents to charge the supply caps. 40.000 uF will probably lead to 40-50A peak charge current, BUT only for 5 or 10% of the time, and then only every other cycle. What you need to look at is the peak current of those IXYS's, not the 6A. That's overkill anyway.
Note that very large supply caps 1) stress your bridge unduly, b) leads to very high and sharp current pulses that generate lots of harmonics to the line freq that sprays all over your amp and 3) doesn't lower the hum in your amp anyway.
Have fun,
Jan Didden
The reason why they use those 25A bridges is because the peak currents to charge the supply caps. 40.000 uF will probably lead to 40-50A peak charge current, BUT only for 5 or 10% of the time, and then only every other cycle. What you need to look at is the peak current of those IXYS's, not the 6A. That's overkill anyway.
Note that very large supply caps 1) stress your bridge unduly, b) leads to very high and sharp current pulses that generate lots of harmonics to the line freq that sprays all over your amp and 3) doesn't lower the hum in your amp anyway.
Have fun,
Jan Didden
Static oddness
The reason I started this thread was that my modded amp started putting out a mid-high frequency "crackling" on deep bass notes when the volume knob hits an abrupt point. It sounds like a burst of static that only last for the brief duration of the bass note.
The amp (200x2) is beefy and it doesn't sound like normal clipping to me.
The only place I could point a finger was at my 6A per diode, handbuilt bridge rectifier. I'm guessing that with 40,000mf to supply they are choking on deep bass notes.
The reason I started this thread was that my modded amp started putting out a mid-high frequency "crackling" on deep bass notes when the volume knob hits an abrupt point. It sounds like a burst of static that only last for the brief duration of the bass note.
The amp (200x2) is beefy and it doesn't sound like normal clipping to me.
The only place I could point a finger was at my 6A per diode, handbuilt bridge rectifier. I'm guessing that with 40,000mf to supply they are choking on deep bass notes.
Most likely not. Remember: the diodes only conduct for about 10% to 20% of the time, when the mains transformer secondary AC peaks above the voltage on the supply caps. At that point, the diodes open up and top up the supply caps, until a few mSecs later the AC dips below the cap voltage and the diodes stop conducting.
Now, from this moment, until the AC again peaks above the cap voltage (almost half a cycle later) the amp can only draw current from the caps. For practical purposes, the whole transformer and rectifier is disconnected from the supply by reverse-biased diodes.
With 40.000mF, there should be enough energy for all but the highest level, deepest, sustained notes. Especially the thing you have, that at a specific point of the level it happens, points to a circuit problem. What *could* be the case is that there is an coupling cap somewhere that leaks. Since with low frequencies the voltage across a coupling cap is largest, it may just 'punch through' above a certain level. But without a circuit, it's difficult to pinpoint.
It may be something else, but it is very unlikely the diodes.
Jan Didden
Now, from this moment, until the AC again peaks above the cap voltage (almost half a cycle later) the amp can only draw current from the caps. For practical purposes, the whole transformer and rectifier is disconnected from the supply by reverse-biased diodes.
With 40.000mF, there should be enough energy for all but the highest level, deepest, sustained notes. Especially the thing you have, that at a specific point of the level it happens, points to a circuit problem. What *could* be the case is that there is an coupling cap somewhere that leaks. Since with low frequencies the voltage across a coupling cap is largest, it may just 'punch through' above a certain level. But without a circuit, it's difficult to pinpoint.
It may be something else, but it is very unlikely the diodes.
Jan Didden
re:
Interesting.....
Your description of how the diode bridge works makes sense, I agree they're probably not the problem now.
I'm using efficient (96db) speakers and probably only have it 3/4 of the way up when the static occurs, but it's only on long, deep notes, as though a power reserve is being depleted.
Look like I have some detective work ahead of me. Thanks...
Interesting.....
Your description of how the diode bridge works makes sense, I agree they're probably not the problem now.
I'm using efficient (96db) speakers and probably only have it 3/4 of the way up when the static occurs, but it's only on long, deep notes, as though a power reserve is being depleted.
Look like I have some detective work ahead of me. Thanks...
re:
They're IXYS Fred...soft-recovery I think. I'll post the part#s tonight, I'm interested in comments.
-Bryan
They're IXYS Fred...soft-recovery I think. I'll post the part#s tonight, I'm interested in comments.
-Bryan
infinia said:
Hi,
How would you suggest that this leads to something that turns up only at low frequencies and only at certain levels?
Jan Didden
janneman said:
In a class B power amp, the average current drawn from each +/- power supply is exactly the peak load current divided by PI (3.14). For 100W in 8ohm this is 5A/3.14=1.7A while in 4ohm is about 3.4A. On top of that comes the output devices bias current (making for class AB, can be significant) and the currents required by the input stage, VAS, drivers (which are usually negligible).
Re: Static oddness
1. The amp could be breaking into oscillation. That would show up on high level transients, and would be more apparent with bass (because HF content would tend to mask it). What that would have to do with the rectifier is a mystery.
2. The rectifier may be miswired, or perhaps has already had one leg go out from charging currents. Check to see if the ripple frequency is 60Hz or 120. If something is wrong and you're only getting half-wave rectification the power supply would get seriosuly overloaded on every bass hit, potentially causing all sorts of trouble.
reiver said:
1. The amp could be breaking into oscillation. That would show up on high level transients, and would be more apparent with bass (because HF content would tend to mask it). What that would have to do with the rectifier is a mystery.
2. The rectifier may be miswired, or perhaps has already had one leg go out from charging currents. Check to see if the ripple frequency is 60Hz or 120. If something is wrong and you're only getting half-wave rectification the power supply would get seriosuly overloaded on every bass hit, potentially causing all sorts of trouble.
If the OP puts his amp back to stock and it fixes his problem. I'll explain what I'm talking about.😀 \\\
hint V=L*dI/dt
fast recovery diodes dI/dt is large and grows with Iload.
hint V=L*dI/dt
fast recovery diodes dI/dt is large and grows with Iload.
Since this only happens with bass and at high volume it could be the speaker(s). I know nothing of your speakers, but I have heard the noise you describe caused by a voice coil coming unglued. Another possibility is that the voice coil is bottoming out or has in the past and damaged the VC former.
- Status
- Not open for further replies.