Hi,
I used this on my first class D design and it worked. But since I intend to build me an ampliverter (HF transformer coupled class D amplifier), I will not be needing this. Ampliverter has no stored energy on secondary and can not produce DC under failure, so I will not need any external protection.
Best regards,
Jaka Racman
I used this on my first class D design and it worked. But since I intend to build me an ampliverter (HF transformer coupled class D amplifier), I will not be needing this. Ampliverter has no stored energy on secondary and can not produce DC under failure, so I will not need any external protection.
Best regards,
Jaka Racman
OPS protection schemes
Hi Jaka (and others):
I see your OPS protection scheme: it's always fascinating for me to see how different minds in different places attack a given problem. In the attached schematic for the PowerDAC 2 output stage, you'll see the route I took.
Now bear in mind that my power rail goes from 2 to 60V to control volume (for volume settings 53 to 99), so whatever it is must work over this range.
At the bottom of the circuit you'll see the dc protection circuit. Strictly speaking it's not a dc protection circuit, but a switching monitor circuit, because in a digital modulated switching amp the probability that the output switching may die is far greater than the probability of a large dc offset being output while switching is still OK.
The circuit consists of a f/f arrangement toggled by the phase nodes of the OPS bridge. The constant current diodes accomodate the 2-60V supply range. The supervisory CPU uses the signals from the dc monitoring to trigger internal counters - if toggling of the phase nodes stops, these counter/s reach a timeout value and the CPU shuts down the entire amp power supply. It's no good simply closing the gate drive, because cessation of switching would likely involv damage to one or other component and the gate shutdown would likely be unresponsive.
The fast overcurrent sensing mechanism consists of an hf current transformer (100t on a 10.4mm 3F3 core; s.o.b. to wind) which triggers a fast comparator. The PWM generator in the CPLD does a latched shutdown of all gate drives in the event that the overcurrent signal is received. Max response time is approx 100ns, from current detect to gate voltage removed.
In a full bridge driven in bipolar mode (Class AD) current flows both in and out of the PSU on each cycle, and the CT won't saturate. What's nice about this hi-side current sense is that it can respond to shorts between LS connections, LS and rails, and phase nodes to each other and to rails. So it's great during prototyping when you're bumbling around with a scope probe in nervous, shaking hands. I've not lost a FET since I started using it, but intend replacing it in future with a GMR current sensor such as those made by NVE because the h.f. CT is such a cow to make.
Cheers
John H
Hi Jaka (and others):
I see your OPS protection scheme: it's always fascinating for me to see how different minds in different places attack a given problem. In the attached schematic for the PowerDAC 2 output stage, you'll see the route I took.
Now bear in mind that my power rail goes from 2 to 60V to control volume (for volume settings 53 to 99), so whatever it is must work over this range.
At the bottom of the circuit you'll see the dc protection circuit. Strictly speaking it's not a dc protection circuit, but a switching monitor circuit, because in a digital modulated switching amp the probability that the output switching may die is far greater than the probability of a large dc offset being output while switching is still OK.
The circuit consists of a f/f arrangement toggled by the phase nodes of the OPS bridge. The constant current diodes accomodate the 2-60V supply range. The supervisory CPU uses the signals from the dc monitoring to trigger internal counters - if toggling of the phase nodes stops, these counter/s reach a timeout value and the CPU shuts down the entire amp power supply. It's no good simply closing the gate drive, because cessation of switching would likely involv damage to one or other component and the gate shutdown would likely be unresponsive.
The fast overcurrent sensing mechanism consists of an hf current transformer (100t on a 10.4mm 3F3 core; s.o.b. to wind) which triggers a fast comparator. The PWM generator in the CPLD does a latched shutdown of all gate drives in the event that the overcurrent signal is received. Max response time is approx 100ns, from current detect to gate voltage removed.
In a full bridge driven in bipolar mode (Class AD) current flows both in and out of the PSU on each cycle, and the CT won't saturate. What's nice about this hi-side current sense is that it can respond to shorts between LS connections, LS and rails, and phase nodes to each other and to rails. So it's great during prototyping when you're bumbling around with a scope probe in nervous, shaking hands. I've not lost a FET since I started using it, but intend replacing it in future with a GMR current sensor such as those made by NVE because the h.f. CT is such a cow to make.
Cheers
John H
hey guys,
just a quick question. I'm thinking about bypassing the input caps on my ucd400 amp (with ad8620); this should be safe as my pre doesn't have much of a dc offset on its outputs.
I'm wondering whether leaving the caps in place and merely bypassing them would sound different than completely removing them.
Any ideas?
just a quick question. I'm thinking about bypassing the input caps on my ucd400 amp (with ad8620); this should be safe as my pre doesn't have much of a dc offset on its outputs.
I'm wondering whether leaving the caps in place and merely bypassing them would sound different than completely removing them.
Any ideas?
matjans said:
Well, this translates in "will a piece of wire soldered without going through the holes in the PCB sound better than one that is going through the holes?" Debatable, shouldn't make any difference in sound quality.
Remove the caps (add solder, heat both legs at one time - cap should come out quick without a hitch), use a soldering pump to clean the holes, then clean residue with a cotton swab dipped in solvent, and solder in a (bent-to-shape non-ferrous) 0.25W resistor leg for the bridge. Looks cleaner this way than bridging between the cap legs, not that it will sound better.
lucpes, that's not what i meant, i was wondering whether leaving the cap in place and bypassing it would sound differently than altogether removing the cap and replacing it with a piece of wire. For testing purposes this is interesting (to me).
If it does sound a lot better without the cap, i will remove it. If not, I'll leave it in place, remove the wire and i'll still have ~original modules.
Although I'm not a novice to soldering or diy, thank you for the soldering advice; it shows you're willing to help other people.
If it does sound a lot better without the cap, i will remove it. If not, I'll leave it in place, remove the wire and i'll still have ~original modules.
Although I'm not a novice to soldering or diy, thank you for the soldering advice; it shows you're willing to help other people.
@Matti
Hi,
If your source has output caps, they will provide a save dc level, so you can bypass them. Better is to drop by and optain a pair of nice ceramic 1u and 22u smd caps to test with. Still use only one pair of caps somewhere in the chain instead of endlessly decouple with rinky dinky caps.
Frank
Hi,
If your source has output caps, they will provide a save dc level, so you can bypass them. Better is to drop by and optain a pair of nice ceramic 1u and 22u smd caps to test with. Still use only one pair of caps somewhere in the chain instead of endlessly decouple with rinky dinky caps.
Frank
I feel like this thread has gone into more advanced aspects of the amps, but I have a novice question about actual building...
I just "finished" building my UCD 400 based stereo amp, but found something troubling about the power supply voltage...
Everything is wired up and between the + and - DC rails I am getting a steady 119.7v on my DMM. I'm thinking to myself "great! 60v +/-"
However when I put the DMM between the + rail and ground, I get 45v+. When I put it between the negative and ground, I get 75v-. How could this happen? What would cause this? Is this a problem? Where should I look to fix it?
When I first powered up the power supply, I had a short, could I maybe have blown a Rectifier diode? Would that do this? I have been using an inrush limiter and, when there was a short, the 10A circuit breaker on the amp tripped immediately.
Just for background, due to what was available surplus, I've ended up with a pretty big supply: 2 identical 43v 725VA toroidal transformers (1450VA total). 2 30,000uf Sprague Powerlytic capacitors per rail (120,000uf total). 2 25A 47uf film bypassed bridge rectifiers. I know it's a bit of overkill, but the total cost was ~ $100 US and I thought that I could always add additional channels later.
Thanks in advance for any recommendations people can give me. Since this is my first amp, I am probably asking really simple questions, so I appreciate the patience.
Thanks,
Larry
I just "finished" building my UCD 400 based stereo amp, but found something troubling about the power supply voltage...
Everything is wired up and between the + and - DC rails I am getting a steady 119.7v on my DMM. I'm thinking to myself "great! 60v +/-"
However when I put the DMM between the + rail and ground, I get 45v+. When I put it between the negative and ground, I get 75v-. How could this happen? What would cause this? Is this a problem? Where should I look to fix it?
When I first powered up the power supply, I had a short, could I maybe have blown a Rectifier diode? Would that do this? I have been using an inrush limiter and, when there was a short, the 10A circuit breaker on the amp tripped immediately.
Just for background, due to what was available surplus, I've ended up with a pretty big supply: 2 identical 43v 725VA toroidal transformers (1450VA total). 2 30,000uf Sprague Powerlytic capacitors per rail (120,000uf total). 2 25A 47uf film bypassed bridge rectifiers. I know it's a bit of overkill, but the total cost was ~ $100 US and I thought that I could always add additional channels later.
Thanks in advance for any recommendations people can give me. Since this is my first amp, I am probably asking really simple questions, so I appreciate the patience.
Thanks,
Larry
lne937s said:
This is for sure not good. I advise to check the power supply first without any amps connected to reduce the risk of blowing up more parts than needed. Are you sure both transformers give the same output voltage? Be carefull with the power supply caps, what is their voltage rating? They should not be overloaded too much. If those big guys explode, you have a real mess., be careful.
Good luck
Gertjan
ceramic capacitors
Hi Frankh
I've been searching the web but am unable to find good quality ceramic caps of the values you mention in your'e post.
With good quality i mean COG or NPO types with a voltage rating of at least 10V.
This is known as a type 1 capacitor. Medium or worse quallity are available (X7R,X5R,Y5V) but there capacitance changes dramaticly with temperature and aplied voltage.
Hi Frankh
I've been searching the web but am unable to find good quality ceramic caps of the values you mention in your'e post.
With good quality i mean COG or NPO types with a voltage rating of at least 10V.
This is known as a type 1 capacitor. Medium or worse quallity are available (X7R,X5R,Y5V) but there capacitance changes dramaticly with temperature and aplied voltage.
UcD400 questions
Jan-Peter or Bruno,
Just how important is the ceramic insulator material you use on the outputs? I removed the heatsinks and one of them fell apart. I was trying to avoid damage to the circuit boards during trial fit and managed to break D24 as well. It appears to be a conventional diode but is it ultra fast or anything special? This I can find and replace but not the insulator. I plan to super glue it together so I can proceed but will need a replacement. I have UcD400AD S# 415 and 416, glad to see they are selling. You guys deserve success!
Thank you,
Roger
Jan-Peter or Bruno,
Just how important is the ceramic insulator material you use on the outputs? I removed the heatsinks and one of them fell apart. I was trying to avoid damage to the circuit boards during trial fit and managed to break D24 as well. It appears to be a conventional diode but is it ultra fast or anything special? This I can find and replace but not the insulator. I plan to super glue it together so I can proceed but will need a replacement. I have UcD400AD S# 415 and 416, glad to see they are selling. You guys deserve success!
Thank you,
Roger
This question is related to other post I made in "Mosfet reliability in class-d amplifiers" and a problem I have experienced with a DIY design.
Is the heatsink of Ucd180/400 connected to power supply's GND in order to avoid that it capacitively couples to the switching node? or is it connected to GND via a resistor or capacitor?
This would give me some clues about my problem.
Thanks!
Is the heatsink of Ucd180/400 connected to power supply's GND in order to avoid that it capacitively couples to the switching node? or is it connected to GND via a resistor or capacitor?
This would give me some clues about my problem.
Thanks!