Powering LM 3886 from Headphone Output, MAC. problem

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Powering LM 3886 from Headphone Output, Mac. problem

Hi I have built 4 LM 3886 amps so far and have no problems with any of them when powered thru my Dynaco PAT 4 or as the output stage for my Dynaco SCA 50 [last project]. But when I hooked one up, which had over a month's use, to my iMac Headphone via an FM 5 thru it's AUX input the amp just quit. It ran well at all power levels then quit after about an hour.
Any Ideas, do I have to put a "real" pre-amp in the circuit? Or is there an easier way to do it, don't really want to use the PAT 4 that I am upgrading to sell or build another pre-amp.

I have included some pix of the SCA 50 [date on pix is wrong, too lazy to reset date on camera]
Used orig Dynyo trans and used the center tape to split the volatage to +- 32v.
non inverting, used the pc boards from chip amp site, twin 21.5 v toriod, snubbed power supply but added a 10k uf/63 to each side [total 20kuf on each side]. Used 100uf on board and added 1000uf in parallel to each side. Used the Dynaco stile output cap on each side [3300uf into 6ft of 20 ga wraped arround the cap and parralells with 10 ohm 3 watt resister, o.1uf mylar in series with 0.47 ohn 2 watt resisters.
No volume control [am using the test amp I built now with 20k pot]
AC or DC Coupled, don't really know, on this amp the ac is not grounded.
Hi had time to do some testing and found that two of the neg side diodes on the power supply blew [over heated?????, added another fan just in case].

Would love to be able to show the schematic but cannot seem to attach it where can I store my images so I can add them??????????????

I am going to try to upload schematic again found a free file host.

And here is a pic of the SCA-50 Project that I just finished remember that the date is wrong on the pic, to lazy to change date on camera

How hot do your "heatsinks" get when it is/was running? The fan doesn't look like it will do much of anything since it is blowing between the two "heatsinks" and not directly on either one. Be sure if you are adding fans to have the fan pointed at the aluminum plate to more effectively cool it.

Also, what is the power rating of your bridge? If it blew, then something is not right (obviously!). If only half of the bridge blew, that lends a little to the DC offset argument (but I am not quite convinced yet :) )The chips themselves are pretty well protected, so I am guessing they didn't blow.

I'll second that your schematic looks a little off. Double check your feedback loop. I am guessing this is a drawing issue and not an issue with your circuit.

Finally, if you make small, compressed images, you can attach them directly to your posts without having a 3rd party hosting.
Your bridge probably blew because of the extra capacitance you have used. If i read you right, you have 20,000uF of supply capacitance!

Personally I would use a 35A bridge bolted to the metal case rather than the discrete schottky diodes, but that's me.
As you have a dual rail supply, +35v and -35v. the speaker output capacitor is completely unneccesary and you can remove it. It's only used in single-rail designs. The original Dynaco was a single rail design and thats why it had an output capacitor. The capacitor will only be adding distortion, especially if it's the original from the Dynaco which will be really old.

You should have a 20K resistor to ground after the 1K resistor connected to the + input to set the input impedance. You should also have a capacitor, say 3.3uF before the 1K resistor to block any DC from the source. Personally I would change the 20K feedback resistor to 22K just for ease, and use the same 22K resistance to ground. This is what I have in my own setup.

Your output coil should be 22SWG enamelled wire wrapped around the 10 ohm resistor you have at the output - this should be at least a 2W device. I used enamelled wire that I pinched from a yellow toroidal choke in a dead PC power supply.

As I said in my last post, your bridge diodes probably blew from trying to charge the 20,000uF of power supply capacitance. What diodes are you using ? IMO, get a 100V 35A bridge like a GBPC3501. They are cheap, powerful and can be bolted to the case. You will never have a problem with the bridge not coping then.
jaycee said:
Personally I would change the 20K feedback resistor to 22K just for ease, and use the same 22K resistance to ground. This is what I have in my own setup.
Unless I misunderstand what you are suggesting, that will set the chip to unity gain, which is NOT a stable setting for this series--the data sheet suggests a minimum gain of at least 10. Typical, though, is around 20 for many applications, and should run just fine with no "real" preamp.

Thus, the revised schematic looks right in terms of the input and feedback sections. If there is DC offset, then a good cap before the 1K input resistor will help, but an output cap is NOT needed.
This is the arrangement I am using in my own amp:

An externally hosted image should be here but it was not working when we last tested it.

Signal ground and power ground are connected at one point only, with a 10 ohm resistor. The output zobel network is 10R + 100nF and is soldered across the speaker output posts.

In case you are wondering about C15, that forms a low pass filter with R1 which rolls off high frequencies that are out of the audio band. This helps to prevent high frequency hash entering the amp and causing oscillation. You can omit it with no serious effects though.

The MUTE pin also has no capacitor to give a switch on delay - in my amp, there is a DC detection circuit seperate that disconnects the speakers for 5 seconds, so it is not needed.

Not shown is the power supply board which has 2x4700uF on each rail and also includes the fuses per channel. It looks like this

I have the amp apart now and will make the changes that were recommended, :)
Probably over did it one the PS Caps, think I will change to bigger 10kuf and remove the second set.
I think I got a bit confused by the single rail schematic from the NSC site and as it mimicked the Dynaco I went with it. This amp was going to be a single rail but all I got was massive feedback from the design NSC showed, so I converted it to dual rail. I have tried the single rail again and it still doesn't work so I will just have to sell the toriods I got from ebay back to ebay.

Thought they were a good deal, 230va, 240v to 120v step downs got 5 for $25, I talked to the http://www.toroid-transformer.com/ when I ordered the dual for my next project [PA100 in a ST120 case] and he told me that these would work as 120 to 60v but not to expect more than 100va for power. Seems a bit odd that he said they would work at full power wired backward for the tube pre amp I am building, but I ordered the right trans from Him, john I think, very helpful.

Thanx will keep you posted

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.