 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
I am having a problem with rectifier diodes blowing up on the power supplies for 2 mono SS amps I built recently. The blow-ups are not ocurring all the time. They happen about every 20th time I switch on the power. I am absolutely certain the construction is correct and there are no shorts circuits. I have done everything imaginable to verify that the construction is correct and that there is no assignable cause for this problem. There is simply no explaination I can come up with except that it is possible that the initial current surge exceeds the surge rating of the rectifier diodes.So here is my question: How would one go about measuring the current surge to a rectifier bridge during the first few milliseconds following switch-on. I have access to a digital signal processor that can sample at up to 100 kHz, but I have no idea how to set it up to measure a potentially large current surge. Can I do it with the DSP? Do I need another tool?
Any help would be greatly appreciated! Thanks in advance.
Rick
Follow Ups:
You can use an X2 rated capacitor (about 0.1uF) in series with the resistor (I use a 10 ohm 1 watt) between the hot and neutral mains. This will help slow the inrush current - I do this as described in the AKSA 100wpc amp plans. Makes switch on transients a little nicer to the upstream components. I also prefer not to swith the amp on/off much and leave it on for weeks at a time.
Just get BIGGER DIODES! This would be expected, IF you turned the equipment on just as the AC was reaching its peak. This will only happen sometimes. Measuring 'could' be useful, but bigger diodes can handle this problem. They can handle a huge, single input surge, if you buy the big ones!
You are asking a difficult question. The reason we use better diodes is to reduce RFI usually generated by improper conduction of normal rectifier diodes. This was pointed out to me by Rick Miller,who wrote an article in 'The Audio Amateur' about 10 years ago.
Rick measured the conduction of different diodes with an RF spectrum analyser when they were used in a power supply. He found that all kinds of crap was generated by normal, cheap diodes.
Schottky diodes, soft recovery diodes, and Hexfreds work well. They turn on and off, properly and quickly. However soft recovery diodes and Hexfreds have more forward voltage drop than Schottky diodes. On the other hand Schottky diodes have lower voltage loss while conducting. This is most probably because of a difference in the physics of the devices. Therefore, for higher voltages, we normally use Hexfreds and soft recovery diodes, but for voltages below 50V, it might be better to use Schottky diodes for slightly higher efficiency.
Say, 20A rating. I simulated your power supply, and the initial inrush peaks are about 20A.Otherwise, install a delayed power up circuit...
Norman
Well... basicaly what you need to measure is peek current. Some better multimeters can do this. I do not know what your DSP is made for, nor what to measure with it. But I assyme it can sample signal from voltage source. So you need a shunt to put in series with output from the bridge and then sample voltage from the shunt. Or you can use existing leads/wires/printed trace to sample voltage. Use closest point to the bridge, and other end of the same wire/lead, or whatever it is. So you'll be measuring voltage drop in this line/wire. Then compare peek value to normal value (that you will confirm by meter), find ratio between two and you'll get actual peek current.
To avoid all this just get bigger bridge/diodes, or get two in parallel.
lead that connects to one of the AC terminals of your rectifier bridge. Use something like a 50 milliohm, high power resistor. Float your oscilloscope and measure the voltage peak across the resistor when you power up the amp. Set your trace sweep to something relatively slow so you can see the current decay as your input caps charge up. You should be able to capture the peak currents after turning on numerous times.If you are using toroidal transformers instead of E-I core transformers, you should note that toroids tend to remain at the point on the B-H curve where you switched them off much more than the E-I cores. The reason I point this out is if you subsequently hit the next turn on near the correct point of the AC cycle that produced the remanent flux in the toroid, you will get a HUGE current spike as the core continues right up the B-H curve into saturation. Something to look for.
Are you using toroidal transformers?
What is the continuous current rating of your bridge rectifiers? Non-repetitive surge current? And PIV?
How much capacitance do you have after the bridge rectifier?
G.
Crank it up...
Sorry I should have mentioned that I'm very much a newbie in the electronics arena, so I will need a bit more handholding if you don't mind. I get the general idea but need more info. Questions/answers below...> Put a small sense resistor in series with ther transformer lead that connects to one of the AC terminals of your rectifier bridge.
What is a sense resistor? Can you specify what I need and where I'd get one?
> Use something like a 50 milliohm, high power resistor. Float your oscilloscope and measure the voltage peak across the resistor when you power up the amp.
I can probably borrow an oscilloscope and I have this digital signal processor that can sample analog input voltages at a rate up to 100 khz, so I should be able to use it to capture the voltage peak/decay. Do I need an oscilloscope or will the DSP do? Also what does it mean to float the oscilloscope?
> Set your trace sweep to something relatively slow so you can see the current decay as your input caps charge up. You should be able to capture the peak currents after turning on numerous times.
What would be relatively slow? Is this just a trial an error process?
> If you are using toroidal transformers instead of E-I core transformers, you should note that toroids tend to remain at the point on the B-H curve where you switched them off much more than the E-I cores. The reason I point this out is if you subsequently hit the next turn on near the correct point of the AC cycle that produced the remanent flux in the toroid, you will get a HUGE current spike as the core continues right up the B-H curve into saturation. Something to look for.
I have no idea what a B-H curve is? Where can I learn about that?
> Are you using toroidal transformers?
Yes, Avel Lindberg Y236751 rated for 30-0-30 VAC secondary voltage. I desire 35VAC output, but had to go to lower rated toroids due to high mains voltage of 125VAC (continuous).
> What is the continuous current rating of your bridge rectifiers? Non-repetitive surge current? And PIV?
The diodes are BYQ28E double diodes rated for 5A continuous (per side) and surge current of 55A. What is PIV?
> How much capacitance do you have after the bridge rectifier?
There is roughly 9400uF capacitance on each side of the power supply.
Thanks very much for your help. I look forward to more info. Let me know if you'd rather to do this via direct email correspondence or if you'd rather just keep posting back and forth here.
Thanks again,
Rick
Hello RickHere are some answers to your questions:
> Put a small sense resistor in series with ther transformer lead that connects to one of the AC terminals of your rectifier bridge.What is a sense resistor? Can you specify what I need and where I'd get one?
Answer: You can get a suitable sense resistor from places like Digi-Key, I would pick something like part # 15FR050-ND that you can see on the following pages of their catalog. It will handle 10 A steady state and much larger peak currents, there are lower values as well that will handle even larger currents.
http://info.digikey.com/T022/V5/0639-0642.pdf> Use something like a 50 milliohm, high power resistor. Float your oscilloscope and measure the voltage peak across the resistor when you power up the amp.
I can probably borrow an oscilloscope and I have this digital signal processor that can sample analog input voltages at a rate up to 100 khz, so I should be able to use it to capture the voltage peak/decay. Do I need an oscilloscope or will the DSP do? Also what does it mean to float the oscilloscope?
Answer: What it means is to put a cheater on the 3 pronged AC plug of the oscilloscope so you can make measurements which are not ground referenced easily. There are ways of doing this on a scope without using the cheater but are harder to describe how to do. This is the simplest. After floating the scope and installing the sense resistor just before the bridge rectifier, put th eground clip of the scope on the end of the 50 milliohm sense resistor that is connected to the bridge rectifier and the probe tip to the other side.
If you use the DSP that you have, you will probably have to re-reference the range that the transformer output to that which the DSP will tolerate which can be done but is likely to be a hassle. Try and borrow the oscilloscope if you can.
> Set your trace sweep to something relatively slow so you can see the current decay as your input caps charge up. You should be able to capture the peak currents after turning on numerous times.
What would be relatively slow? Is this just a trial an error process?
Answer: You want to capture multiple line cycles so you can see the initial pulse and subseqent charging cycles as you turn on the amp. Since each 60Hx cycle is 16.6 milliseconds and a 50Hz cycle is 20 msec, I would set the scope to no less than 20 msec/div and more likely to 50 msec/div on the horizontal scale. Yes this is somewhat a trial and error process. The slower trace rate will give you better persistence on the scope screen so you will be able to see the peak current better.
I skipped a couple of your answers in this space, thanks for the info it helps with what follows:
> What is the continuous current rating of your bridge rectifiers? Non-repetitive surge current? And PIV?The diodes are BYQ28E double diodes rated for 5A continuous (per side) and surge current of 55A. What is PIV?
PIV is Peak Inverse Voltage which is 150V for the BYQ28E that you are using. This should be fine given your rectified DC voltage of ~+/-45 VDC which you are probably getting. However if you have used reasonably large wiring on your primary which I suspect you are, you are getting well over 10A peak inrush current to charge up your 9400uF input capacitors. These diodes are only rated for 10A repetitive peak. Even under normal operation you are likely to get that. I suspect your problem is the choice of these diodes. I recommend that you try to get the BYQ30E as soon as possible. These are a direct drop in and give you 16A continuous and repetitive peak and a non repetitive peak of 80A. Still not enough in my opinion but it will make an improvement.
(These diodes you have chosen are intended for high frequency switchmode applications where they see a square wave input from the switching transformer and relatively lower output capacitance than you have chosen. Therefore they see much lower peak to average rectified current ratios than diodes that are used for 50/60 Hz AC line rectification. I know these fast recovery types are in vogue for tweaking amplifiers but are in general not as rugged as conventional bridge rectifiers. I am trying to get you info on more conventional rectifiers but the OnSemi web site is currently down for maintenance)
We can take this to email if you need additional info. It is a hassle cutting and pasting into this dialog box they give you.
Regards
Joe...
G.
Crank it up...
It is more rugged than any of the others I listed earlier. The OnSemi web page is back up. These are a direct drop in to your circuit as well.http://www.onsemi.com/pub/Collateral/MUR1620CT-D.PDF
OnSemi (used to be Motorola) makes the best power semis in the industry IMHO.
These devices have a non repetitive surge of 100A vs the 80 for the BYQ30E that I mentioned earlier. I feel that these will solve your startup problems.
Joe...
G.
Crank it up...
This post is made possible by the generous support of people like you and our sponsors:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2025, Audio Asylum, all rights reserved ]
