Right now I have two incomplete Phonoclone kits. That's all I have at home right now. Depending on how much money I'll get from Rhys I'll maybe buy the parts myself.
. Let me know how it goes then.
Unfortunately not. Rhys was on his way to the post office when I called him (he was outside at least) but some made that the package/letter never got sent. It should have been at my place last Friday. I have tried to call him every day since a few days back and the phone is on but he doesn't answer. A few minutes ago the phone was switched off (batteries could be discharged also). When you have a mobile phone you can't say that you haven't seen 5-10 missed calls so he is fully aware of I have called. It's possible that there is a possible explanation.
I'm in process of building 4 PA150. I connected to the power supply. 3 boards remained cool. one fried! I don't know at the moment what caused it but the second you connect it to power one of the LMs is smoking while the others are EXTREMELY hot. I'm trying to take the smoking one out but it is far from simple. I will have to have his pins cut. Luckily I ordered more than the 12 required to complete the project. However, I don't have a clue yet to the cause of this.
I cannot spot any shorts to ground on the output.
After the offending LM is taken out, the other two are still getting warmer than those in the other boards (somewhat hard to compare though, because all other LMs are attached to a massive heatsink while the examined LMs only reside near it).
I now have two things to consider:
was the offending LM faulty from packing or is there something in the PCB (other component, hidden short) that wrecked it?
have the other two co-LMs suffered damage due to the ill behaving one and should be also replaced (hard work and potential damage to the PCB)?
your thought and ideas are welcomed.
My advice here is to omit the output resistors and test all the three LM3886 unconnected. And after power up measure the output offset voltage. The real cautious person should solder one LM3886 at the time and test each one. It's extremely important to make sure the gain is exactly the same for all three LM3886. My guess is that you may have missed to place the passives correctly.
How do they get hot? Do you have currents in the series resistors?
How do they get hot? Do you have currents in the series resistors?
P-A,
After removing the burned LM, I run the circuit with only two LMs and even brought the CD offset of one of them to 0V using the trimmer. however, being worried by the rising temperature (one can touch the chip but I did not want to wait to the point it gets hot beyond touching - lost my nerves here!), I switched the PS off after that.
you are absolutely right about working step by step. However, as one who is coming from the software world, and being impatient, I used the "success oriented" approach. in software you burn no components, only your time in searching for a bug you just wrote. obviously here this approach does not apply. I did my best to compare the placement of the passives to the scheme, mask and other PCBs. at least all resistors are in correct places. the capacitors are impossible to identify so I don't know.
I don't know how to tell if the other LMs are damaged. the only evidence I have that something is abnormal is the temperature. I also somewhat reluctant to solder the output resistors as it took a fair amount of heat to melt the solder beneath it. I need a way to make sure the circuit is working without connecting it
After removing the burned LM, I run the circuit with only two LMs and even brought the CD offset of one of them to 0V using the trimmer. however, being worried by the rising temperature (one can touch the chip but I did not want to wait to the point it gets hot beyond touching - lost my nerves here!), I switched the PS off after that.
you are absolutely right about working step by step. However, as one who is coming from the software world, and being impatient, I used the "success oriented" approach. in software you burn no components, only your time in searching for a bug you just wrote. obviously here this approach does not apply. I did my best to compare the placement of the passives to the scheme, mask and other PCBs. at least all resistors are in correct places. the capacitors are impossible to identify so I don't know.
I don't know how to tell if the other LMs are damaged. the only evidence I have that something is abnormal is the temperature. I also somewhat reluctant to solder the output resistors as it took a fair amount of heat to melt the solder beneath it. I need a way to make sure the circuit is working without connecting it
I gathered my wits, replaced the faulty LM and after meticulously checking placed the PA150 back on power. it now function well in the sense no smoke is coming out and heating is similar to the other board (hot, though). trying to kill the DC offset, I found myself running out of resistance and still have 6-20mv on the various LMs.
P-A, you write it is important to match the LMs' gain in each triplet, but the only control I have is over the dc offset. the gain is fixed by the fixed resistors.
P-A, you write it is important to match the LMs' gain in each triplet, but the only control I have is over the dc offset. the gain is fixed by the fixed resistors.
The gain is one thing. The other thing to trim away the offset with the pot but I'll guess you may experience perhaps more offset in rare occasions since the offset voltage can be 0-10 mV times the gain. 20 mV offset at the output is good but the current flowing between the chips is the offset divided by you series resistance and this is a compromise. High resistor value, lower heat, higher output impedance, hotter resistors at high output power.
DRV134 hiss
getting one step further, I assembled a DRV134 circuit (chip, 2 capacitors, connectors) on a strip board. as others before me, I encountered an excessive hiss on the drv outputs. I suspect it is circuit layout dependent, but I don't know how to verify or contradict it. I need to test the drv board in a more insulated environment.
I also suffer from another dreaded power amp problem - subtle hum. however, this hum seems to be picked up by the input cable, as it is there even when the cable is not connected on the other end.
getting one step further, I assembled a DRV134 circuit (chip, 2 capacitors, connectors) on a strip board. as others before me, I encountered an excessive hiss on the drv outputs. I suspect it is circuit layout dependent, but I don't know how to verify or contradict it. I need to test the drv board in a more insulated environment.
I also suffer from another dreaded power amp problem - subtle hum. however, this hum seems to be picked up by the input cable, as it is there even when the cable is not connected on the other end.
P-A,
do you mean cabling between the RCA plug/PA150 and the circuit? they are all shielded and shield acts as ground. the noise is present only when the drv134 is powered. when fully connected (power, in and out) but not powered, or when I remove the circuit and let the wires to the PA150 hang free on air on one side there is no noise. when pulling the input to ground using 1K resistor the noise is almost totally muted. however, checking with an oscilloscope, I suspect the noise on the individual phases is about the same as without the resistor, but less correlated, and hence does not cancels out when bridging.
have you completed checking your implementation of a DRV board? do you have any noise issues?
do you mean cabling between the RCA plug/PA150 and the circuit? they are all shielded and shield acts as ground. the noise is present only when the drv134 is powered. when fully connected (power, in and out) but not powered, or when I remove the circuit and let the wires to the PA150 hang free on air on one side there is no noise. when pulling the input to ground using 1K resistor the noise is almost totally muted. however, checking with an oscilloscope, I suspect the noise on the individual phases is about the same as without the resistor, but less correlated, and hence does not cancels out when bridging.
have you completed checking your implementation of a DRV board? do you have any noise issues?
for PS I'm using TeddyReg from PinkFishMedia :
http://www.pinkfishmedia.net/forum/showthread.php?t=39990 with added LED on each rail.
the drv implementation has only connectors, and 2 electrolytic capacitors in parallel to 2 film capacitors between the +-V and ground for smoothing whatever is caught on the PS relatively long lines. I've tried adding a pulldown 100k resistor on the input with no change in behavior. also tried 1k resistors between outputs and ground - nothing changed.
The main abnormality I can spot is a difference in the level of noise (and DC offset) between normal and inverted phase pins. I cannot find an external cause for that difference.
any wild ideas are welcomed.
Shmulik
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.