18V bipolar power supply for LM3886 amp

[PRR]

Congrats on getting the 3886 working.

What R.G. said. Or IOW....

I do not understand the tube heater wiring.

It looks like you'll have 4V heat on the 6V heater connection. It sure won't be happy that way.

It looks like you have monster common-mode ripple on your heater leads (very close to sensitive audio leads).

I do not see how you can be getting even 10V DC at the input to the "+12V DC" system.

The tube plates sure will love 28V DC.

Tubes do not need regulated power.

The 100K-250K resistors in the tube and tone-stack will NOT drive a 10K pot well.

I do not think you need a volume control between tone-stack and '3886.

The under-volted heaters will reduce tube gain severely. The 10K load throws-away another 10:1 gain. Input sensitivity could be over 140mV; hard work. Between under-heating, likely low B+ (lower even than 12V which is already pathetic), and serious 10K loading, the preamp probably can't hit the '3886 strong enough to make it work hard, even half-hard.

As R.G. said, take 1K and 100uFd from +28 to the tube B+ line.

Throw-away the 6V rectifier and regulator. Take 6VAC direct from transformer, through leads tight-twisted right up to the tube socket. No 8 ohm resistor. (-OR- use a dedicated 9VDC wallwart WITH the 8 ohm resistor.)

Throw-away the 10K pot. Add 470K from '3886 input to '3886 ground (mostly to hold it if you disconnect the preamp). Connect '3886 input to treble pot wiper. There is an issue here: the pot will scratch and there will be part-Volt DC offset through to the speaker. But it will work a whole lot better, maybe no problem.

Input sensitivity in this configuration should be near 4mV, plenty for any non-thrash work.

[El Heisenberg]

Gunna post a pic later tonight. I had mounted the 100uf cap backwards!!


I didn't want to run the filaments off the 12v regulator cos I wanted the big glow, but if that's what it comes down to i'll do it. I was already planning on fiddling with the anode supply voltages but I was just stumped that the amp was so quiet.


RG...um what kind of a buffer?

[El Heisenberg]

Ok wutever RG and PRR says I'm doin. 

[El Heisenberg]

PRR, when I just connect pin 9 and pins 4 and 5 to the ac rails, I measure 13.2v AC when I measure pins 4 and 5 to pin 9. With the resistor it measures 12.6v, or 6.3v depending on how I take the measurement.


Actually, with the transformer I'm using now, it's only 1.2A. And I'm getting different measurements. So instead of the 8 ohm resistor, I'm using a 2.2 ohm resistor, and it gets me 12.1vAC from 4,5 to 9....or was I supposed to get 6.3 from 4,5 to pin 9...I can't remember...maybe that's why I used the 8 ohm. I actually didn't have an 8 ohm resistor, I settles on three 2.2 ohms in series.

[PRR]

> I didn't want to run the filaments off the 12v regulator cos I wanted the big glow

The heaters MUST be fed the RIGHT voltage.

A 12AX7 heater can be wired two ways. For 6.3V (5.7V-6.9V) or for 12.6V (11.4V-13.8V). Under-volt a little, the amp sucks. Over-volt a little, and it will work OK for hundreds of hours rather than thousands of hours. Yes, people do go outside the specified ranges. But get it working first, then play games.

> wutever RG and PRR says I'm doin.

There's differences. Pick one.

R.G. suggested changing the 10K to 1Meg followed by a buffer. This is Good Design. Buffer could be TL071 etc.... but there's no elegant way to power it here. A JFET would be dandy, but it needs a resistor and a cap and a resistor and it is way too humid here tonight to be doing drawings.

I say you do not "need" a gain trim here (there's one after the first triode). And that you "can" connect '3886 input to the treble pot wiper.

My plan is cheap-trick. The tone-stack has zero to 510K of DC resistance. The '3886 has 0.2uA-1uA input current. There could be up to a half-Volt of induced DC (but typically 0.1V DC). The '3886 is wired with a cap in the NFB loop, so input DC is passed to the output with unity gain. Half-Volt of DC will not annoy any loudspeaker suitable for 50 Watts of Guitar. It will "work". You will be able to HEAR yourself without arm-sweat, and you will be able to diagnose the hum issues.

[El Heisenberg]

Yes, I already ditched the volume pot.

[El Heisenberg]

This helped marginally with the hum::























Maybe I should do two stages of those? Or maybe different values? That's all I had on hand. I can make trips to the shack or go tearing up some VCRs tho... I GOTTA get this thing done!


This amp is loud enough to play with a drummer. I GOTTA finish it!!

Also, it's on the breadboard so that may be causing some hum. The pot shells aren't grounded either. When I touch them I get a big hum.

[El Heisenberg]

HAHAHAH! This amp is rediculously loud!

[El Heisenberg]

Hey, why not run the tube plates with 56v???


The 1k resistor and 2200 uf cap got rid of alot of noise. I think wutever noise I was getting from the plates is taken care of now. But I don't know, maybe different values will get rid of more noise. The amp is usable now at least. But the hum is still really present. I think it's comming from somewhere else. If I turn the volume pot of the first tube stage all the way down, I still get the same amount of noise. If I ground the input of the amp, it's totally totally silent.  

I kinda think I might want a second volume control before the power amp cos the tube preamp sounds different with the volume low.






any problemos here?

[El Heisenberg]

nevermind. trying to run the tubes at 56vDC and still have bipolar supply to the amp was a dumb idea.

[PRR]

> it's on the breadboard so that may be causing some hum. The pot shells aren't grounded either

A decent gitar amp has enough raw gain to pick up ALL the hum in the room.

You breadboard to get it working. Yes, this should be a LOUD amp.

Ass-ume most hum now is wide-open hay-wiring. When 90% "done", you will move it inside a low-hum room: a metal chassis. With input far from output and farther from power supply. With effective grounding. I suggest you hard-ground the input jack, buss along the stages and pots to the speaker-common isolated from chassis, then on to the power supply which is not tied to chassis. Your wall-cord ground wire does go to chassis.

> 1k resistor and 2200 uf cap

That's the 30-pound hammer. Use a 2-stage filter. The tube stages pull about 0.25mA, it would be reasonable to allow 10% drop in voltage for 99% drop in hum. Rounding, this is two 5K (4K7) resistors and a couple 10uFd-100uFd 35V caps.

Aside from lower cap-cost, the several-stage filter does a far better job of reducing the high harmonics of 120Hz ripple.

> the tube preamp sounds different with the volume low.

Then go back to the level pot shown on '3886 app-sheet but make it 250K or 500K.

Does the treble-pot "scratch"? The DC input current of the '3886 is "poor design", but 1uA bias won't hurt the pot, and any "scratch" may be mild.

[El Heisenberg]

? So the only part of the entire circuit that is grounded to the chassi is the input jack?  

Sorry If I'm misreading  or something.

I was going to tie the power supplys ground lead to the chassis earth post along with the power cord ground.  I also heard it was good to attatch the speakers ground close to the bridge rectifier.

[petemoore]

  that site...
  About the grounds...can't beat reads on grounds for ground questions.
  There are so many it gets confusing.
  Speaker AC currents through the ground wire put a signal through the wire [outside it too but don't run the input wire along it and it won't make a 'transformer = inductance of two wires run along each other]...sensative wires get sensative 'real estate', heavy load wires get their own 'sewer pipe' to the 'dump'.
  A star ground system is nice for proto-work, otherwise if you can figure out how to save copper with X Y Q or K layouts [pretty much would have to design the amp with copper savings up on the priority list, knowing how by trying layouts]...or read about what you can ''get away'' with. Usually the input jack and input volume [example] can share the same ground wire...small currents of same source.
  Other shared grounds may or may not introduce noise, Starring them of course is a good starting point since it's generally low noise regardless, there may be caveat/arguements about lowest possible noise.
  I think the best way to figure out the 'trick' grounding is to read and make 1/2 trick, or for full tricking [least copper, quickest const.]:
  Have a control amp handy that is very stable and low noise.
  Try to beat or cheat it.
  Beat: increase of stability or trick a drop in the noise floor.
  Cheat: less copper/wires and still stable/low noise.
  Sometimes it's 'that wierd stuff' [instability] that you don't even really know is there that's hardest to get rid of.
  Sometimes 'mixed sewage' [from star ground..a long wire..ground Y like input common Gnd. to jack and input volume control] or sometimes everyting to do with a tube stage grounding] doesn't send a pulse/signal ''back-up the ground line'', sometimes it does.
 

[El Heisenberg]

PRR, just fired it up. The treble pot is not scratchy. The volume pot after the first triode hums and bumps loudy when I use it. Whenever I touch it it goes "WUUUUM". But it doesn't scratch when I turn it. Still got other noise too.

petemoore, the sites pages that you directed me to are where I'm raising my questions from. I read up all I could before I even attatched the chip to the heatsink. I thought I was stuck so I just read up and then sat around for 2 weeks.

I still gotta get a silicon or rubber washer and a thermal pad or something.

[El Heisenberg]

Ugh drove to 3 radioshacks today looking for a product. Their website says it's stocked in the stores, but they're wrong! Looks like I ain't the only one liftin stuff from the shack.

But I finally found ONE heatsink mounting kit they have for TO-220. So i got a mica washer and a thin little clear slip. The chip is wider than the slip tho, but the metal part runs down the middle and it SORTA lines up. The top could still make contact tho. But I decided to try it anyway cos I have nothing else and it can't hurt. The heatsink measures 0 volts. Tomorrow I'm gunna find a second one. It'll be a long drive in the hot hot sun. I just wanna have a second one and cut it up and stick it there to be safe. I forgot to put thermal paste too. It's not conducting as well. The chip gets hot at full power and doesn't radiate the heat away to the heatsink as well with the insulating slip. Hope that's just cos no paste. Once I get that done I'm gunna bolt the heatsink to the chassi. I dunno what I'm doing. I'm just guna drill pilot holes, then get a screw and hope it goes in. What I REALLY wanna do is just put electrical tape all over the heatsink and use rubber feet and stuff to keep the keatsink from touching the chassi. Seems harder and easier to do at the same time.



nice thumb shot

[El Heisenberg]

[Brymus]

Looks cool, how does it sound aside from being loud ?
Thats the preamp PRR designed right ?
Radio Shack has the thermal grease too,kinda pricey but you dont need alot.
I did the same as you at first ,bought two TO220 mounting kits and cut the mica from one to fill in the gap from the first.
I later bought some real mica sheets from All Electronics,as well as getting some what look like vinyl insulators from the supplier I bought my LM3886 chips from,he says they are for insulating them but they look like flat ribbon cable without wire in them.
I didnt get any further as I wanted to perfect my tube preamp, then got caught up in other projects 
If I were to do it over again I would spend the extra couple of dollars and get the LM3886T which doesnt need insulators.
I have since bought a toriod capable of driving two 3886's in bridge formation to get a true 100 watts but still other stuff keeps taking my time.

[El Heisenberg]

Welll I have pictures of the enlosure before and after. I was going to come on here and post them all along with a long post on the problems.


There is still hum dammit. I orginally had two plastic jacks. But either the two I was using from an older amp were broken, or they just didn't work....there....I think they were broken. I put in a metal one and the amp worked but there was hum. I covered both sides of the hole with duct tape, making sure to get the inside rim. The hum went down. I can now turn down the preamp volume and get total silence. So the amp section is quiet. There's still hum from the preamp.

Another strange thing. I've tried two different tubes in. One of them would flash really brightly when I flipped the on switch. It was startling.

I'll post pics tonight and everything with the wiring. I star grounded. I used the guidelines from RGs article and his advice in this post. I'm gunna try to switch with the jacks. Put the metal one at the input. I still haven't measured the filaments either.

The chip is insulated tho. It works. But if I woulda known that the LM3886T didn't need that I woulda just gotten that!!

The hum comes when I turn the bass knob up. With the treble all the way up and the mid and bass all the way down there's only slight hum. Also there's hum when you turn the volume down. There's the least hum with the volume all the way up.


Aside from that I really like it. I like the way it sounds with just the treble knob on 10. It'd be nice to be able to use it other ways. But it's just the PERFECT volume to play drums with. I'm so happy with it. If only I can get this hum out. The hum isn't a big issue on the setting I keep it on. But with the mid or bass up it get's bad.

[El Heisenberg]

here. I made the schem show how I grouped the grounds together and head of the ground symbols is a wire to a long screw and a nut I put in. The mains earth wire is at a different point.

I tried to follow everything everyone's said here and everything here::

http://www.geofex.com/Article_Folders/stargnd/stargnd.htm

but I'm still getting hum. The noise is from the tube preamp cos it only happens after the filaments heat up. Originally I had a wire going from the speaker return to the 0v point at the capacitor leads. After putting this wire to the star ground the hum from turning down the volume went away. But I'm still getting hum from bass and mid control. A tiny amount and some other negligable noise with the treb at full.



When I put my hand near the tube or cup it, the noise is totally gone.

Oh wait, I see.  I need a tube sheild to stop this don't I??? Oh man. OOOOOOh MAan. Just when I thought I was done! Arrrgh AAAArgh! I was celebrating.

[El Heisenberg]

The chassis is a VCR/DVD player. The bottom of it anyway. I put wood sides on it. Spray painted it black. Then mounted the transformers and the tube socket. THe transformers were heavy so I put a peice of wood so didn't buckle and look odd. I used those I dunno what you call em. Terminal strips or something. I used some euro style terminals for the captive mains cable cause I didn't want to cut out a hole for an IEC socket. And I didn't wanna find an IEC socket either.

I'm not real big on planning. So all my builds turn out messy. Don't laugh. I'm working with junk! But please point out everything I did wrong!

I couldn't bolt the heatsink to the enclosure. I ended up drilling a hole through one of the fins and putting a screw the the metal lip through the wood and into the fin. It's stuck there. You can see it in one of the pics

the extra pot is for a digital reverb with the belton reverb brick I'm gunna put in.






































Guess now I gotta find a tube shield. And a label maker. I ain't using my usual white-out pen for this thing.


I hope the tube shield fixes everything. Thanks everyone who helped me and were so patient with me. Thank you. This thing ROCKS, if it were without hum...







UPDATE: I wrapped the tube in tin foil. Ugly. Can't see the glow .

Got rid of ALOT of the hum. Still some when I turn the bass up... dunno why.


But it's ALOT better now. ALOT better. Dammit are there tube shields with slits so you can see the glow? And what's the remaining hum comming from?