Important Question about Power Supplies and PTs

[smackoj]

I have just about completed a 68 watt Gainclone amp and I'm stuck on the power supply. The Chipamp manual for the LM3886 amp show on Page 4 a schemo
of the Power Supply. They are using a PT that does not have a Center Tap but the PT I have DOES have a CT. Is there a way I can still use this PT with a CT or should I just buy a different PT?

Thanks, smacko
 

http://chipamp.com/         Sorry, I couldn't get the page of the manual to transfer. It's under "Manuals"  "LM3886 Kit Manual"   Page 4 Power Supply Scemo

[sluckey]

Is there a way I can still use this PT with a CT...

Sure. Using a center tapped secondary will even make the circuit simpler. I'm assuming the secondary voltage is appropriate for your circuit?

[smackoj]

Great, thanks Sluckey.  Yes, I have 24 volts AC and the manual states that's a good number before recto and big E caps.

adios, 

[sluckey]

A 24VAC with center tap (same as 12-0-12) will give you +17VDC and -17VDC at the filter caps in the circuit I drew.

[shooter]

On page 11 it talks about the secondary volts and effects different volts have.  I just used a SS PA in my last build.  It wanted about the same as yours.  I used a "junk" tranny outta a SS fender bassman 20-0-20
the output power is lower than spec'd but good for my app.  looks like from the page you want 18-24vdc per side on the caps.

[PRR]

The manual wants *dual* 18VAC-25VAC secondaries.

As Slucky says, a single CT winding is perfectly acceptable (the dual supply is some kind of audiofile silliness), but logically a single winding should be twice each side of the dual winding.

You want 36VCT-48VCT.

Go ahead with the 24VAC CT for testing.

The 2*24V or 48VCT would give about 35V each side, 70V total. Counting on thumbs this will cover about 24V Sine RMS, which in 8 Ohms is 72 Watts.

The 24VCT will give 17V each side, 34V total, about 12V Sine RMS output, which in 8 Ohms is just 18 Watts!!

You are getting a quarter of the power possible with the chip.

You "wasted your money on that 68 Watt chip", working it nearer 18 Watts. (However the chip is cheap, hardly any more than an 18W chip.)

Testing at quarter power is probably a fine idea. Get your wiring figured out, play pretty loud, it will be fine.

But the right transformer is a *big* 2*24V or 48VCT. Like 48VCT @2A or 100VA.

[smackoj]

Sorry gents but I guess I didn't describe my PT correctly. It's a big PT from a Peavey Renown SS amp that has 32-0-32 vac and I am using a 1200 ohm 5 watt power resistor on each side to get apprx. 24 vac on each lead. I thought from reading the manual that 24 volts prior to the "snubber" recto and filter pcb was close to the max power the writer believes makes this chip work best?

I see now from the schemo Sluckey sent that I will have to change the recto diode layout because the pcb for the power is designed for a transformer without a center tap. I don't think I will be able to "cut and paste" on this "Chipamp" pcb so I'll have to do it some other way. 

I'm open to suggestions if anyone has one? I have a really stout regulated power supply that has + and -12 vdc at 2.5 amps. Does it make more sense to buy a couple of the chips that raises the voltage? I don't know if they make them that can go from 12 to 30+ volts? 

gracias amigos.   

[PRR]

> 32-0-32 vac and I am using a 1200 ohm ...to get apprx. 24 vac on each

So you have 8V drop in 1,200 Ohms, you must be drawing 0.006,7 Amps.

When the power amp is LOUD it will want to suck about 1.8 Amps. 200 times as much!

What will really happen is the power rail will drop below 1V each side, and power output will be near zero (very distorted, if you get anythng at all).

In a transistor amp there can be a HUGE increase in power supply current from idle to roar. Far more than most any tube amp. So dropping resistors will cause HUGE SAG, typically to near-nothing.

> 32-0-32 vac

So 64VAC total, which will cap-out to 90.5V DC. The limit on LM3886 is 94V MAX. You are way too close to the limit, IMHO.

Note that over-volting a tube amp leads to short life. Over-volting a transistor amp can be instant death.

> chips that raises the voltage?

That's a round-a-bout approach. Generally a Power Amp should directly get raw right-voltage power.

While there may be a kludge, GainClone transformers are readily available on eBay.

As for your Reknown PT... I think it is just a bit more than a LM3886 can safely take long-term. Maybe if you put a 12V buck transformer in front, and use a BIG heatsink. Even then I fear the Reknown PT is just too much for the puny LM3886.

[shooter]

the 1st paragraph in your manual, pg 11 says a DIY user is using a 30volts without problems.  Also calls for a 2 or 3a slo blow fuse.  I'm thinking you could try your tranny with a zener on the CT.  I attached what is used in tube amps, just add the zeners to Sluckeys drawing and measure your power supply WITHOUT connecting it to the amp board.  I would hope/shoot for +/_ 35 to 40vdc at the points that go to the amp.  More than that n you are probably entering the danger zone.  Also, a 16ohm speaker but don't really what range impedance the amp can use.

 EDIT:  the zeners in the schematic are probably not the right value for your application, I think that configuration drops 40ish volts.  It COULD? be a cheap option if it works n you can figure out the value of zener

[sluckey]

I would not try using zeners in that configuration with a dual polarity power supply.

[smackoj]

I must be on the wrong thought track with this Peavey PT. I thought I could use 32 vac dropped to 24 for the Plus rail and the other 32 vac for the negative rail which would give me the 35=40 vdc on the plus and the minus sides after rectification?

Is this what I need to make this power supply pcb work? The pcb works as shown in the Chipamp LM3886 Kit Manual (schemo on pg. 4 of pdf)

http://www.mcmelectronics.com/product/287-1055

 

[shooter]

It appears you have a separate PS board, I would wire it up - stand-alone - not hooked to the amp board, plug it in and see what your rail to rail volts is, If it's above 80 (+/_ 40) then I'd look for another option.  PRR explained the droping R's, If your supply sags that bad, your amp will probably sound that bad. I don't think it would damage the PA, so like most here, if we're pretty sure it won't break it, try it.   I'm running mine (20-0-20) but there is a definite lower wattage, sounds closer to my 18W tube amps than my 35w. 

Sluckey, I wasn't sure about the Zeners, seemed logical, is there a explainable reason on dual supply?  thx

[smackoj]

Shooter; thx. Yes, it is a stand alone PS pcb that looks exactly like the one on the Chipamp.com website. I will lite it up without connections to the amp first, but I'm still uncertain what transformer to use? Is this what I need?

http://www.newark.com/triad-magnetics/vpt48-520/transformer-toroidal-2-x-24v-25va/dp/11R8079

thanks much,  smacko

[shooter]

Right volts, wrong current.  The manual suggests a 220VA rating, the one you referenced is 25VA.  I can't find anywhere what the amp draws for current.  We can "ballpark" it.  calls for a 3A on the ac side, 120. the secondary is 50.  That's a 2.4:1 step down.  so current will step-up, so 2.4*3 = 7.2A. secondary.  Hopefully one of the sharper knives will verify or clarify that.
I'll try and hunt down the tranny that was recommended for my SS amp because I'm pretty sure it will work in your build.  It was pricey, like $60ish

[shooter]

Here's the best I can find.  did a google search on 220VA 30-0-30 and wound up in china!  But;
http://www.tnt-audio.com/clinica/ssps2_e.html
That will take you to a page explaining way more than you probably wanna know about PS's
http://www.digikey.com/product-detail/en/80-2/80-2-ND/1984785
That will take you to about the closest thing I can find in tranny's (without getting insane on price) it's a 160VA which your datasheet indicates will work with some loss in power. 
My build was a VM100 SS amp that spec'd a TR8040 tranny which is only a 100VA

Good luck

[smackoj]

thanks Shooter. good info.

jack     

[PRR]

> what I need to make this power supply pcb work?

eBay is full of the right parts. Here's a search to get you started:

http://www.ebay.com/sch/i.html?_from=R40&_nkw=LM3886+transformer

Not all will be just-right.

The Signal part is an older-line US maker and will do fine, after you look-up the actual connection diagrams.

Toroid Transformer T140 is a commodity part; *maybe* "better" than the China sellers because it is in Hong Kong (however HK is China so it may be a front).

In both (and all) cases the Shipping is about as much as the iron.

Are you in the US? Does anybody remember the name of the transformer shop that is in the same building as ParMetal in NJ?

[MakerDP]

Does anybody remember the name of the transformer shop that is in the same building as ParMetal in NJ?

Antek. It was suggested to me recently that this transformer from them would make a fine Gainclone trannie...

http://www.antekinc.com/as-3225-300va-25v-transformer/

I am about to make the plunge and do a Gainclone myself and I will probably use this transformer or probably the 22v version actually.

[PRR]

> Antek

Thanks for name and link.

For a single guitar channel (or a stereo hi-fi), he only needs like 150VA. Antek's listing is spotty, here's a suitable 200VA which is in-stock at the moment:

AN-2222 - 200VA 22V Transformer $29.00
(USPS Priority to Maine $12.00)
Grand Total: $41.00

[MakerDP]

Yeah a 200VA should be fine. Usually 80VA per channel is the recommended bare minumum.

[smackoj]

Yes, "Antek" is one of the outlets I found when searching for Toroid type transformers.  I didn't realize when I got started on this project that there was a whole other world of transformers with "dual" primaries and secondaries. I had to do some quick reading to understand the VA rating and the fact that the Chipamp power supply pcb has to have 4 leads going in to make the pos. and neg. rails. It's all new stuff to me. This kind of shoots my idea to build a powerful but inexpensive amp. I knew I had lots of transformers laying about the shop and I thought I could put one back into service. Now I realize that to use the pcbs I already purchased I have to buy another transformer. I guess it's not really a bad thing when you figure if I was going to college to learn this stuff I would be 20 or 30 grand in debt for student loans.

thanks to all who replied.

smack jack

[MakerDP]

Any reason you couldn't use two transformers if you had two suitable single-secondaries laying around in your stash?

[smackoj]

Maker; yes, I tried that idea but all my trans. have a center tap and the power supply pcb is only configured for trans. W/O a CT. Darn. So, it's either a toroid type or build a different power supply circuit and I'm not confident on doing that unless someone can offer a proven design that won't fry the board, the cat and burn down the kitchen. As PRR has pointed out, this PS has to be rock solid or the whole project is headed for melt down.

 

[shooter]

You can avoid burning up the cat but if by chance you do, I have 11 you can choose from!
Look at page 4 of your manual, the schematic Sluckey posted and the one I posted.  The +pgnd and -pgnd are both Common, or ground to the board - I think.  You can verify that by looking at the foil traces or ohming those 2 points, verifying they are shorted together.
Now from a "standard tranny" solder 2 wires to each end, and 1 to the CT and wire like my drawing.  The CT wire can be run to any convenient point along +Pgnd or -Pgnd.

You may wind up with the suggested tranny and that's cool, it's just there is more than one way to skin a cat!

[PRR]

> wire like my drawing. 

I *suspect* that plan is flawed (a cat-burner). However I am fighting an internet problem and have not had enuff coffee to sort a new LinkSys, much less a power supply.

[smackoj]

thanks PRR. laughter is good medicine.

 

[PRR]

Argh!! The company plan has a TYPO on it!!

The lower bridge rectifier (FWB) is drawn Wrong-Polarity!

See top plan in attached image.

I ass*ume the actual PCB is correct.(?)

You really want one FWB.

I *think* you can use the PCB verbatim by removing (even ignoring) the extra diodes and doing some jumpering. Middle plan. However it makes my head hurt.

What *I* would do is use the PCB to hold the caps. Then get a big fat FWB assembly and bolt it to the chassis (power supplies this big usually benefit from rectifier cooling). Wire the PT to the FWB's "AC" (or "~") terminals. Wire the FWB's "+" and "-" terminals, and the PT CT, to the PCB cap traces (perhaps to the holes where you took-out their rectifiers).

While the PCB shows two "GND" traces, which *may* be gooder for a dual-secondary design, you have just one lead on the PT. So jumper the GND traces at the output, and also *AT* the big capacitor "GND" terminals. I assume the amplifier PCB has both GND pads strapped together; in that case you only need one wire from PS GND to Amp GND. 

[shooter]

Sorry I killed your cat!, I was thinking it might not work with2 rectifiers after the post.

[smackoj]

thanks amigos.  I have a nice grasa FWB recto unit that should be just the ticket. I'll give it a try and report back.

 

[smackoj]

Finally, a good report. The FWB device is wired as per diagram no. 3 above and it's putting out a rock solid 34.8 volts plus and minus dc. I'm stoked to get the little amp chip hooked up and see what kinda mojo it makes?  Further reports to follow. 

thanks guys. 

 

[PRR]

> a rock solid 34.8 volts plus and minus dc.

?? Is this still the "big PT from a Peavey Renown ... 32-0-32 vac"? And you have it wired rectifier to capacitors? Then it should be closer to 44V DC each side.

If no caps, rectified un-filtered DC computes as 29V but 34V might happen no-load and high wall voltage.

If this is some 24V+24V transformer, 35V on the caps is as-expected.

Can't be rock-solid under heavy load. (OK, there's different kinds of rock.)

+/-35v at idle is a good spot for LM3886. 70V is safely below the 84V (94V no-signal) MAX rating of the chip. If the +/-35V were granite-solid, you'd be on the vertical dashed line below-- 63 Watts in 8 Ohms and over 120W in 4 Ohms (except the chip won't sustain 120W out). If I make a reasonable assumption about the PT (it is firm clay, not granite) then maybe the slanted purple line-- 45W into 8 Ohms, 72W into 4 Ohms, both perfectly good for the chip (WITH ample heatsink- but it will shut-down if you do get it too hot, and your finger is your clue). These are low-THD numbers... in gitar-amp world you can probably add 30%-40% no-lie, 60W and 100W.

[smackoj]

I guess I am "not sure" about the rock solid?  I thought the dc volts could be low figuring the PT makes just under 32 vac from my wall socket (yes, Peavey Renown PT). I took the reading from the + and - lead to the CT ground not a chassis ground. I didn't have it mounted in the chassis yet. Maybe that gave me a skewed number? 

I have not hooked it to the power amp pcb yet.....good thing fo sho.

jack

[shooter]

Not sure if it's apples to apples but my supply loaded came in lower than expected by the math even though it's stable, looks good scoped and metered.

[sluckey]

I guess I am "not sure" about the rock solid?

Rock solid means the + and - VDC will not vary from no load to full load.

DON'T USE THAT PEAVEY 32-0-32 TRANSFORMER! It will put out +45vdc and -45vdc which is too close (may even exceed) to the max rating for the LM3886. And don't think about dropping the voltage using resistors because that would introduce sag which will destroy the "rock solid" supply. See this datasheet...

     http://www.ti.com/lit/ds/symlink/lm3886.pdf

You are way over thinking this power supply. The Hammond 167P50 rated for 25-0-25 @ 5A is a good match for the LM3886. It will produce a rock solid +35VDC which will give you 50W Cont. Avg. Output Power into 8Ω, even more power into 4Ω. (see the datasheet again)

The attached circuit is all you need to safely power that LM3886. Be sure to use a good heatsink and silicon grease. Simply bolting this chip to an aluminum chassis is not sufficient for high power.

[smackoj]

sorry 'bout the "rock solid" tong slip. I thought it looked good but I see your point about 'too close' to the meltdown line. I really don't want that to happen.

I have a big heat sink and some tranny grease so I'm ok there. I'll have a look at the transformer you suggest.

thanks mucho. 

 

[PRR]

> It will put out +45vdc and -45vdc

90V total.

> which is too close

LM3886 is rated 84V with signal, 94V NO-signal.

Agree this transformer is too much voltage for LM3886.

Several alternates have been proposed in this thread.

I do not know any easy-to-use chip which will take the Reknown's high voltage. That was a big amp, bigger than chip-amps. (Also perhaps "bigger" than the Reknown's discrete design could stand?)

[smackoj]

"too close to the line"  yes, I get that now. I did run the amp last night with a Stancor PT that makes 14vac on each side (20vdc after the FWB recto). It runs good with very nice articulation on the guitar notes BUT it gets hot muy pronto so I only ran it a couple minutes (bread boarded sorta). I am going to look for a more suitable PT and then mount it up on the heat sink with some grease.

I hope to add a few pics and a sound file asap.

 

[6G6]

On a somewhat related note, I have whats left of an Acoustic B220.
The P/S, preamp, chassis and case are all there, working and great looking.
I have used this as a standalone pre amp.
The PA section is gone.
Previous owner gave it to me and said that board was toast, so I threw it out.
No repair on that one!
Rails from P/S are + 54 and -54.
Original output was 220W @ (I'm guessing) 4 ohms.
Could I get away with paralleling 4 gain clones by using the +V and gnd
on half and -V and ground on the other?
Meaning not really ground, but the 0V point on the P/S.

[shooter]

Are you talking stacking NPN/PNP pairs in the PA section?  I only have a couple schemos and they are only single stacked pair, nothing in || but I know it's done in split-rail PS.  worked on a few that used 20 each in || per rail.  I loved it when one shorted, then I could use my steel fuse n find it quick!!!!

[6G6]

I am talking about trying to power the LM 3886s by using + 54 on the + side and 0V on the - side to be well under the max voltage,
while still using the P/S I have.
Then using 0V on the plus side and -54 on the - side on another pair.
Bottom line is, so far I no $ in this and would like to stay close to it.
Found gain clone kits from Hong Kong on Flea Bay @ $7.48. 
Maybe I am getting carried away with wishful thinking.
If it happens, I'd have a spare bass amp I could leave at my practice place, instead of an extra door stop in my shop/storage building.
Now that I think about it, I remember seeing a ground connection on the op amps.
Maybe I could run V+ to one, connect V- leg to V+ leg of the other and V- to second op amp's V- leg, to split the voltage.
 
 

[PRR]

> Rails from P/S are + 54 and -54.

There's no good way to run chips from that.

Indeed you "could", by wiring each chip for single-ended supply with an output capacitor. A very BIG output cap. At least twice (one for each rail), and probably four times (because the two chips on each rail won't hit the same DC level). Cost and bulk are significant, and it is conceptually ugly.

> said that board was toast

Possibly worth investigating (except you tossed it). Amps this big are a b*tch to trouble-shoot, but a once-working amp is a better bet than cobbling together a bunch of less-capable parts.

[6G6]

I was afraid it wasn't going to be that easy.
If I had been able to get my hands on the original PA board,
I bet I could have patched it up.
It's not like I have never soldered a wire jumper across a burned up trace.
Alas, I never even saw it.
Un common amp to begin with and the only place that even claimed to have a schematic wanted $25 for it.
No biggie, in that I have $0 in it and wasn't actually in need of it to begin with.

[PRR]

Yuan-Jing Mono Audio Power Amplifier Board 1 x 350W
http://www.parts-express.com/yuan-jing-mono-audio-power-amplifier-board-1-x-350w--320-6520

This *may* work with that power supply.

Documentation is nil.

The claimed 350W@4 is not possible on the claimed +/-45V supply. That would need +/-52V, which is near what you observe. I'm sure your +/-54V sags at full power to more like +/-44V to give the apparent 220 Watts.

It *appears* to have the right stuff for a >220W@4 amplifier. The output devices will stand 140V, so 108V leaves some margin. 

It lacks a heat-sink, which must be BIG. Skimpy sinking will lead to melt-down at full roar. 

ANY high-power amp build is tricky and can become costly. Any slip-up burns amny costly parts.