I found that leakage through the supplied 100uF cap matched the current fed through the 220k and refused to let the base voltage rise above 500mV. The relay would not turn on. Adding 100k//220k solved this problem.
I also found that the 639 was faulty. Changed to a bc550b with the legs bent to fit the pad layout.
I have not yet tried re-forming the electrolytics before installation. This may be a better solution.
You must insulate the 3886 from ground.
That is why Uriah included the insulating patches. Use them.
I also found that the 639 was faulty. Changed to a bc550b with the legs bent to fit the pad layout.
I have not yet tried re-forming the electrolytics before installation. This may be a better solution.
You must insulate the 3886 from ground.
That is why Uriah included the insulating patches. Use them.
Yes...
try to insulate them and try again but probably you fried the LM3886 and possibly the LM318...it could be R11.
If so buy again those parts (buy the already insulated LM3886TF to avoid further problems)
I wouldn't settle the LM3886 too close to the board, leave at least 1-2 mm from the bending of pins.
Attachments
Last edited:
Simply insulate and try, resettling the LM3886 you'll risk to destroy it...
If you buy new LM3886TF than you can settle them a little higher.
My 2c.
Andrew, are you saying you had TWO faulty parts in your kit? That is most disconcerting. Is it possible soldering overheated the transistor?
I can understand a failure due to an uninsulated chip, but the number of failures seems to be almost equal to the successes with this latest group buy. Is that why so many builders are not reporting?
Is there any way to test these parts before they're put into the circuit? I check the value of all resistors and caps before inserting them. Will merely testing an electrolytic cap help to reform it? What else is there to do?
I dread the thought of building any more duds. My troubleshooting skills are sorely lacking.
Peace,
Tom E
I can understand a failure due to an uninsulated chip, but the number of failures seems to be almost equal to the successes with this latest group buy. Is that why so many builders are not reporting?
Is there any way to test these parts before they're put into the circuit? I check the value of all resistors and caps before inserting them. Will merely testing an electrolytic cap help to reform it? What else is there to do?
I dread the thought of building any more duds. My troubleshooting skills are sorely lacking.
Peace,
Tom E
I can confirm that the only 639 I have soldered in so far had to be replaced. I will not attribute blame.
I can also confirm that the supplied 100uF and the 220k timing resistor are not compatible as supplied.
That does not equate to two faulty components. It took me a long time to get the relay to click over. I too am not good at troubleshooting.
I have already commented on the timer/DC detect on the MyRefC, again I resisted repeating my view on this part of the design.
The only way I can see to make significant improvements to it, is to add too many components to make it easy for all to modify. I intend adding a very small daughter board in lieu of a few components in this area to get very much better performance. Once I have this optimised I should be able to give details if some want them.
I can also confirm that the supplied 100uF and the 220k timing resistor are not compatible as supplied.
That does not equate to two faulty components. It took me a long time to get the relay to click over. I too am not good at troubleshooting.
I have already commented on the timer/DC detect on the MyRefC, again I resisted repeating my view on this part of the design.
The only way I can see to make significant improvements to it, is to add too many components to make it easy for all to modify. I intend adding a very small daughter board in lieu of a few components in this area to get very much better performance. Once I have this optimised I should be able to give details if some want them.
Last edited:
Andrew, what value is your R14? Does any of this have to do with higher voltage to the relay circuit? I ask because, based on Bill P's conjecture and your subsequent discussion about too low voltage there, I reduced the value of R14 to 270 as a sort of compromise. I haven't finished my build yet, so I don't know the consequences.
Peace,
Tom E
Peace,
Tom E
"is that why so many are not reporting"
I can only speak on my own account: Yes!
I have one channel that still doesn't make music.
I have replaced both the chips 3886 and 317, the latter defintly working tested before and after use.
Next step is replacement of the the transistors, they iust arrived. This nuissance plus holidays cause delay in reporting.
My 2cts
Michiel
I can only speak on my own account: Yes!
I have one channel that still doesn't make music.
I have replaced both the chips 3886 and 317, the latter defintly working tested before and after use.
Next step is replacement of the the transistors, they iust arrived. This nuissance plus holidays cause delay in reporting.
My 2cts
Michiel
I am guessing and I hope I am right.
I think the low amount of build related posts is a result of people buying this amp so that they can have it when they want it. They know GB's for it are far and few between and yet it is such a great amp and usually a very easy build. I think many kits are simply not built. Also dont forget how much trouble I had getting kits out in a timely manner. My apologies on that but think that it has only been a short time since my last shipment relative to the number of non audio things going on in out lives.
I am sorry to see another relay not kicking in. I dont know why this happens but it almost always seem to be the LM318. Even in the previous GB the same problem rose its head a few times.
The LM3886 must, without fail, be insulated by both the insulation pad as well as the shoulder washers. If not you will kill it.
Was there a problem with a 220k resistor? What could I have done wrong here?
I have a lot of the matched resistors left. If anyone wants some of those just let me know. If you go buy some LM318 I suggest buying more than one and more than one brand if you are trying to replace one that is acting up. Remember our builder that bought 5 of same brand and not a one worked.
Uriah
I think the low amount of build related posts is a result of people buying this amp so that they can have it when they want it. They know GB's for it are far and few between and yet it is such a great amp and usually a very easy build. I think many kits are simply not built. Also dont forget how much trouble I had getting kits out in a timely manner. My apologies on that but think that it has only been a short time since my last shipment relative to the number of non audio things going on in out lives.
I am sorry to see another relay not kicking in. I dont know why this happens but it almost always seem to be the LM318. Even in the previous GB the same problem rose its head a few times.
The LM3886 must, without fail, be insulated by both the insulation pad as well as the shoulder washers. If not you will kill it.
Was there a problem with a 220k resistor? What could I have done wrong here?
I have a lot of the matched resistors left. If anyone wants some of those just let me know. If you go buy some LM318 I suggest buying more than one and more than one brand if you are trying to replace one that is acting up. Remember our builder that bought 5 of same brand and not a one worked.
Uriah
You did not do anything wrong. The design called for 220K but it seems in a few cases capacitor leakage current steals enough base current from Q2 to prevent the relay from closing. I chose to reduce the value of R14 to make the +24V supply actually be at 24 volts. That helps to push enough current through the 220K to close the relay more reliably. Andrew chose to keep a low supply voltage and reduce the 220K to about 68K and supply more base current that way. I haven't analyzed that approach to see what it does to circuit timing. Either way, the relay circuit design seems marginal with the original component values and a few builders may need to tinker a bit to get it to work reliably.
(btw R11 near 318 is Ok)
FWIW, you can omit the relay circuit entirely - it's independent from the audio signal portion. The turn-on click is barely audible, and there's no turn-off thump either - it's fairly well-behaved even without the relay. I assembled four of the Twisted Pear boards this way, without any issues - however, I'm using the LM3886TF isolated plastic package version and the TI LM318 CERDIP, which I think is better margined.
Yes, you can do that and all is well until one of the LM3886 devices fails. It has the potential to take out your speaker if the LM3886 output device(s) short to a supply rail. I participated in the group buy in part because a decent protection circuit was on the card.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.