Hi Mark,
Thanks for your inputs. Did you know you have X-ray vision? I rechecked the entire PCB and sure enough I did find several un-soldered points. I took me 5 minutes to retouch all solders and make them shine like a new penny but 15 mins. to run them through a high power magnifying glass. And guess what... The amp came to life.
The sad part though is that there's oscillation in the speaker protection relay. It comes on for 1 sec and then goes off for a fraction of a sec and comes back on and repeats this. Only when I touch the top of cap C9 and C24 on the PCB (C9 and C24 on Mauro's rev c pdf) does it stay on. Strange, isn't it?
Now the power I'm giving it is 22.5V (AC). I cut out two 6" pieces of a 1000W water heater element and put them in series with the winding ends. That makes them hot, but I'll have the transformer fixed once I get the amp finalized. If I go below 22.5V (AC), {I gave it 17.3v AC} the relay comes on only when I turn on the amp and then goes off within a sec. and then stays off.
Some questions:
1. What is the ideal AC voltage that I need to supply for driving a 4 ohm load. (My understanding is 22v AC, but I need to confirm). The ebay page says 24-28v but doesn't mention AC or DC. Link: Classic Lines LM3886 Dual Channel LM318 Amp DIY | eBay
2. Oscillations relay: Dealing with a relay based setup for the first time, I don't know why is it oscillating at all. I'm attaching my current wiring setup. The speaker leads are 3m each. The input is about 1m supplied by an android smartphone.
If need be I'm willing to replace the LM3886. Could 29V AC have blown off the LM318n or the speaker protection transistors? The LM3886 heatsink goes hot (untouchable) in about 1 min.
Attaching rev a (for speaker protection) and rev c for reference plus some photographs.
PS: Mark, may I buy you beer/coffee?. You beat the pixel peeping eyes of a professional web designer.
Thanks for your inputs. Did you know you have X-ray vision? I rechecked the entire PCB and sure enough I did find several un-soldered points. I took me 5 minutes to retouch all solders and make them shine like a new penny but 15 mins. to run them through a high power magnifying glass. And guess what... The amp came to life.
The sad part though is that there's oscillation in the speaker protection relay. It comes on for 1 sec and then goes off for a fraction of a sec and comes back on and repeats this. Only when I touch the top of cap C9 and C24 on the PCB (C9 and C24 on Mauro's rev c pdf) does it stay on. Strange, isn't it?
Now the power I'm giving it is 22.5V (AC). I cut out two 6" pieces of a 1000W water heater element and put them in series with the winding ends. That makes them hot, but I'll have the transformer fixed once I get the amp finalized. If I go below 22.5V (AC), {I gave it 17.3v AC} the relay comes on only when I turn on the amp and then goes off within a sec. and then stays off.
Some questions:
1. What is the ideal AC voltage that I need to supply for driving a 4 ohm load. (My understanding is 22v AC, but I need to confirm). The ebay page says 24-28v but doesn't mention AC or DC. Link: Classic Lines LM3886 Dual Channel LM318 Amp DIY | eBay
2. Oscillations relay: Dealing with a relay based setup for the first time, I don't know why is it oscillating at all. I'm attaching my current wiring setup. The speaker leads are 3m each. The input is about 1m supplied by an android smartphone.
If need be I'm willing to replace the LM3886. Could 29V AC have blown off the LM318n or the speaker protection transistors? The LM3886 heatsink goes hot (untouchable) in about 1 min.
Attaching rev a (for speaker protection) and rev c for reference plus some photographs.
PS: Mark, may I buy you beer/coffee?. You beat the pixel peeping eyes of a professional web designer.
Attachments
Thanks. Finally I got it to work and stay on consistently.
1. I raised the voltage till the time the Xfr would give about 24.8V (AC).
2. I measured 28.4v (DC) on Pin 1 of LM3886.
3. I measured (-) 12v on pin 4 and (+) 12v on pin 7 of LM318n.
3. I did put a 4ohm resistor in series with my 4ohm speakers to bring them up to ~8ohms.
4. I hear a lot of high feq. noise as if you are trying to tune into a SW radio station (like the crickets you hear at night).
Questions:
a. Still clueless about offset. How do I measure the offset and what should the ideal value be? It only seems to turn on if I supply it some audio signal. By offset, do you mean the voltage on the relay primary?
b. How loud is it supposed to sound? Like a old valve radio? It sounds too low. (The speakers are not in their cabinet yet.)
c. How soon should I expect the LM3886 to become untouchably hot? Currently it's 45sec with a 10cm x 7cm x 4mm.
d. Do I need to raise the value of C34 from 10pf to 100pf. Some report that this makes it more stable. (sorry I asked).
e. In order to drive 4 ohms, what should the AC supply be?
Can't wait to get this thing inside it's case. Though I'm yet to put up LM1036 controls for it.
1. I raised the voltage till the time the Xfr would give about 24.8V (AC).
2. I measured 28.4v (DC) on Pin 1 of LM3886.
3. I measured (-) 12v on pin 4 and (+) 12v on pin 7 of LM318n.
3. I did put a 4ohm resistor in series with my 4ohm speakers to bring them up to ~8ohms.
4. I hear a lot of high feq. noise as if you are trying to tune into a SW radio station (like the crickets you hear at night).
Questions:
a. Still clueless about offset. How do I measure the offset and what should the ideal value be? It only seems to turn on if I supply it some audio signal. By offset, do you mean the voltage on the relay primary?
b. How loud is it supposed to sound? Like a old valve radio? It sounds too low. (The speakers are not in their cabinet yet.)
c. How soon should I expect the LM3886 to become untouchably hot? Currently it's 45sec with a 10cm x 7cm x 4mm.
d. Do I need to raise the value of C34 from 10pf to 100pf. Some report that this makes it more stable. (sorry I asked).
e. In order to drive 4 ohms, what should the AC supply be?
Can't wait to get this thing inside it's case. Though I'm yet to put up LM1036 controls for it.
Probably still unstable, or the input LPF (3k3/220pF) has a dry joint.
a. Output Offset is the DC voltage at the speaker terminals with no signal, and the protection relay operating normally, i.e. connected. You can measure it with a DC voltmeter between the speaker out terminal, and Power GND on the board. Anything below 100 mV is OK, though you can get down to a few mV in many cases.
b. It depends on your source level - if you have a line-level signal (say from a CDP), it will be very loud indeed - no way to listen to it reasonably without a volume control pot at the input.
c. If it's stable, it should never get untouchably hot even with modest heatsinks. Rapid heating is a key symptom of instability.
d. For a correctly-functioning Rev C, you should never need to increase C34 outside the range of 10 pF to 27pF, and very rarely above the 10 pF default.
e. 18-0-18 to 22-0-22 should be OK for 4-ohm loads, but it will have poorer sonics and higher distortion than 8-ohm loads.
Thanks linuxguru for a point-wise reply. There are no dry joints as of now having retouched every joint with my 20 yr soldering experience and passing each joint under a huge magnifier.
The offset is 5v and 4.2v at the two channels. It only varies from 5.0 to 5.2 and 4.0 to 4.2 with & without input. The amp is being fed by an android smartphone as audio source only at volume from 0 to 50%.
So what now? Blown up ICs? Anything else I can check before replacing the ICs? I'm planning to throw in TL071s for LM318-n and then maybe replace each 3886. (The first time I powered it up using a 30-0-30 Xfr and with several leads dry soldered due to manufacturing defect). Now I'm giving it somewhere ~23/24 - 0 - ~23/24 V (AC) since my mains keeps fluctuating.
The offset is 5v and 4.2v at the two channels. It only varies from 5.0 to 5.2 and 4.0 to 4.2 with & without input. The amp is being fed by an android smartphone as audio source only at volume from 0 to 50%.
So what now? Blown up ICs? Anything else I can check before replacing the ICs? I'm planning to throw in TL071s for LM318-n and then maybe replace each 3886. (The first time I powered it up using a 30-0-30 Xfr and with several leads dry soldered due to manufacturing defect). Now I'm giving it somewhere ~23/24 - 0 - ~23/24 V (AC) since my mains keeps fluctuating.
The offsets are too high, and you'll cook your woofer voice coils pretty soon with those kinds of DC offsets. Measure the offsets again with the source disconnected (i.e. no connection to the inputs). If it's still high, replace the LM318s and proceed (you can use TL071 for testing). Check continuity of R11 after that.
With inputs open the relay oscillates like 4-5 times per sec. The offset therefore is random but I even measured 11v sometime while it was oscillating. The speaker connections were keeping it at 5v and would have been cooked pretty soon.
R11 & R35 seem to be ok. They are 1 ohm in the circuit, my meter reads them 0.
R11 & R35 seem to be ok. They are 1 ohm in the circuit, my meter reads them 0.
It's unstable even at DC, which indicates that there's something wrong in the circuit or the assembly, for sure. A correctly-assembled Rev C is unconditionally stable at DC, with output offsets below 0.1V.
Check everything in your implementation against Mauro's reference schematic(s) at the beginning of this thread. Measure and check the values of all resistors, and all small caps if possible. Replace the LM318 with TL071 for further testing.
Soongsc: TL071 is a good and cheap option to troubleshoot without burning the LM318-n which happens to be kind of rare animal these days.
linuxguru: If I pull out the LM318, then it's stable and the relay comes on. I'm going to check each part with my meter. Though when you say something is wrong, ICs are the first thing that come to my mind. Are you implying that something other than the ICs is bad?
linuxguru: If I pull out the LM318, then it's stable and the relay comes on. I'm going to check each part with my meter. Though when you say something is wrong, ICs are the first thing that come to my mind. Are you implying that something other than the ICs is bad?
I just pulled out the LM318n and put in TL071 and played this guy for over 45 mins. I put in all my favourite songs and every 4-5 mins I blew some air on the heatsinks (just in case).
The power supply is 24Vac and gave me 33.8 Vdc (prone to mains fluctuation). I'm planning to reduce the voltage further since I'm going to drive 4 ohms.
So what now? Change to new LM318n? Or keep TL071?
I really can't thank you guys enough. Your experience and willingness to help and entertain my newbie questions are the reason why this thing works.
The power supply is 24Vac and gave me 33.8 Vdc (prone to mains fluctuation). I'm planning to reduce the voltage further since I'm going to drive 4 ohms.
So what now? Change to new LM318n? Or keep TL071?
I really can't thank you guys enough. Your experience and willingness to help and entertain my newbie questions are the reason why this thing works.
The recommended load for the MyRef is 8 ohms, so you should stay with that if possible. It's more stable and has lower distortion numbers.
If you're OK with the sonics of the TL071, you can stay with it for a while. However, the MyRef had its compensation specifically tailored for the LM318, which is probably the most transparent-sounding opamp in this application, and also faster/higher slew-rate than the TL071.