...again.

About a year or so ago I ordered a chip from National Semiconductor and built my 1st chipamp to use as a practice guitar amp. I followed the schem from the datasheet and guessed at a PS from reading at Rod Elliott's website. I built all this on perf but made a bunch of rookie mistakes cause I was too cheap to buy one of Brian's kit's or boards. A series of calamities culminated in the utter destruction of this 1st amp when the charged heatsink made contact with ground resulting in a very firey smokey end to that 1st chip. I had made the mistake of buying the uninsulated version and then didn't follow good advice to insulate it myself.

So now, I acquired a PCB set from Joe and set about doing this right. I even used the recommended polypropylene caps this time. It sounds really good and doesn't even get warm so far whereas the old one stayed hot due to oscillation caused by my horrible perf layout. I'll try to get some pictures up within the next few days.

The preamp I used was already there and is well documented. It may not be for everyone but I like it a lot.


RDV

I just picked a 20v + 20v @ 5amps(200VA) Talema transformer. It's supposed to be a single primary and dual secondary. I guess it's a few years old NOS so I can't find any info on any website.

There are two double insulated(orange covered by yellow) stranded wires, and there are four solid wires(black, white, red, & green).

Would anyone know the deal with these? Would the two stranded wires be the primary? That's the only thing that makes sense if this is a single primary/dual secondary transformer.

What's strange to me is the lighter(though double insulated) wires being the line voltage & the heavier being the 20v.

I also searched the net for the word Talema and the only diagram I saw was not like this transformer.

I got continuity between the two yellow wires, between the black & white, and between the red & green.

I measured the resistance between the three pairs and got 3.3 ohms between the two yellow wires.

.8 ohms between the other two pairs.

Does this mean the two yellow wires are the primary? I got a response at DIYaudio saying that the larger resistance would be the primary, but I'd like an R.G. opinion on that one.

Please advise.

RDV

I put the Dr. Boogey together tonight in a marathon 5 hour session. It was a bit tedious wiring 6 pots but boy was it worth it. This is the best sounding high-gain DIY distortion yet. I'm not sure you could make it better unless a footswitchable cleaner sound could be made available. With that, you'd never need anything else. It really sounds good through my chipamps and I suspect a clean amp is where this will shine brightest. I always have trouble with super high gain circuits through my Marshall. The timbre of the distortion reminds me of my Fab-Metal but with less fizz and more meat!

A heartfelt DIY thanks to Electrotabs!

RDV

First of all, tranformers gots polarities. See those dots on the transformer schematic? All the leads with dots go positive (and negative, in turn) together.

Each winding used as a primary needs to be driven with its rated voltage. In the case of the dual 115V ones, you connect dot to dot, non-dot to non-dot, effectively paralleling the primaries.

If you get one winding reversed, you will see major smoke when you power it.

The transformer's rating is based on having enough amperes flowing into the core so the secondaries can take their designed current out. Dual primaries are made half the otherwise-required wire size each, so to get full power out, you must have the designed current flowing in each primary winding. You do this by paralleling for 115, or series for 230. Both situations give you equal current in each winding, and contribute each primary's fair share of ampere-turns to the M-field inside.

If you series two windings, you must connect a dot to a non-dot, and then put voltage across the series setup. again, if you connect it backwards, major smoke pours out, and the transformer is itself tranformed into a Darkness Emitting Diode (also called a DED).

You cannot get more power out of a tranformer than it's designed for by any means except running it at a higher frequency. So, no, hooking each secondary up individually will not help. What the spec is telling you is that each 18V secondary will do 1/2 of the specified 18V current when they are paralleled; this is the same current they will supply in series, which is shown as 36Vct. Same voltages, same currents, hooked in series instead of parallel.

You want primaries parallel, secondaries in series for your amp, since the LM3886 is designed for +/- power supplies.

I prefer a 42Vct 2A transformer for the 3886. Parts express used to sell one of these for $15, but I can't find it at their site now. Even better is a 115/115/22/22 toroid. I got a few of these on ebay.
_________________
R. G. Keen

I've got a ground loop occurring with my subamp which I'm running from an aux out on my Mackie mixer, it goes from there to an active crossover, then into a chipamp into the sub. I disconnected the ground on the signal cable that goes from the mixer to the subamp. Doing this took the hum away from my near-fields but sort of increased it in the subamp, well either that or now that it only comes from there that's how I'm perceiving it, I don't know. Would a 1:1 isolation transformer be the way to go, leaving the signal ground unhooked still? I'd like a little advice before proceding.

TIA

RDV

I'm about done with non-insulated versions of chipamps.

I forgot that I had an insulator around my speaker jack for the original perfboard chipamp guitar-amp I built and fried it last night by touching the metal plug casing to to the heatsink and grounding it out. SNAP!! What a sick feeling. I'm gonna insulate them from now on or I'm not gonna even plug them in.

I'm gonna cut the chipamp part of the perf off and use the rectifier section on another chipamp. I'll post anything positive.

RDV

Here's a design I pulled from DIYaudio.com that was running on 9v power. I removed the biasing network and added grounds for bipolar power. Thanks to Brion55. I may test this one myself.

RDV

R.G. Keen was right. Again

He said don't have a charged heatsink, he says use an insulator.

I didn't. A guitar cord that was plugged into my mixer made contact with the charged heatsink and blew the chipamp to smithereens.

This was a lot of work down the toilet. Argh.

It was the amp running my subwoofer. I don't think anything else is hurt. It didn't even blow my fuse.

The smell is atrocious. The chip phisically split in half and smoke was rolling out of it.

F**k!!

RDV

RDV

I've got a preamp in my initial chipamp/guitar amp I built that is from the old Thomas/Vox(USA made) SS amps like the Super Beatle.
I like it, but I think we should come up with a very neutral sort of simple preamp that would have some guitar voiced EQ but little else so pedals would determine the sound a great deal sort of like what R.G. Keen has done with his Workhorse amps. I don't want to copy R.G.'s design or anything because as we know his is an all tube design.

I suppose some kind of limiting circuit might warm things up. The problem as I see with a SS front end in general is that it is too clean. I suppose what I'm trying to get at is I think we need something that is barely on the edge of breakup if you really lay into it, but still retains as much headroom as possible. I like teemuk's ideas about diodes but I don't think I want there to be much available distortion, perhaps not even a pre-volume. I'm just throwing out ideas here because I've not really been satisfied with things I've come up with in the past.

RDV

"I used a small heatsink for the chip, so I had to use the fan."
RDV

"OK, special advice.
As I tell people a lot, power amps are essentially a power supply with some other junk on them that lets a little of the power out carefully. That's not a complete description. Power amps are also heat-exchange engines as well as power supplies.
Before you make yourself a sealed up amp, go to ebay and buy one of the flat-back heatsinks that come up there every now and then. You can get on for under $10 that's heavy with lots of fins. 3" along the fins, 6" wide, and maybe 1.5" to 2" fins would be good. Make that heatsink part of your amp. Ditch the fan. You have the beginnings of a very good amp."
R.G.

"I cranked back up and was blasting away and the chip shutdown on me (for a couple of seconds), the fan doesn't cut it at gig volume. But then R.G. said that didn't he?
It will however work for the low-volume stuff I do at my desk.
I'm ordering a proper heatsink very soon though, and perhaps a chassis."
RDV

"I bought one of these and I thought I'd cut it in half and make 2. There's really not too many fins now that I look at it though. What do you think?"
RDV

"I actually started to snag a couple of those myself. That'll certainly be better. Go for it!!"
R.G.
____________________________________________________________________________________________________________
"I think this might be an ideal transformer for an LM3886, any thoughts? Perhaps the 250va one instead?"
RDV

"Actually, that's good for about two of them.
22+22 volt secondaries is about right, total of 60Vdc, +/-30. That's close to the max you'd want to use with these chips. You should get 50Wrms out of the thing with +/-30 supplies. With 50W out, efficiency of about 79%, you need to supply 50/.79 = 63W of DC from a 60V supply. That's 0.95A of DC. A full wave rectifier makes the transformer see an RMS current of about 1.6 to 1.8 times the DC current, so the transformer is seeing maybe 1.7A. The VA rating is then 44*1.7 = 75Va.
What that tells you is that a 75VA transformer will run one of these things at a full power sine wave forever without overheating.
But music is not static. It has peaks and valleys, and transformers have enough mass to "average" the power demands in terms of their internal heating. Stereo transfomers are commonly sized to supply 1/2 to 2/3 of the continuous power rating, and they don't suffer much. You could probably get by with something under 75VA if you wanted to play brinksman.
Parts express used to sell a 22+22 EI core thing at about 80VA for $15. They were the best match to the LM3886 I ever found. I don't see them there any more."
R.G.

"The subject of how much transformer to buy is a tricky one.
If you want full power at 100% duty cycle, it's easy. You add up the watts, and then compute average DC currents. From that you compute transformer heating current in RMS (they're different!) and pick your transformer VA as the product of its output voltage and the heating current RMS.
That always give a pessimistic answer, because music is not constant power. Only unnaturally compressed sound or constant signals can do full power all the time, and people won't listen to that. Music varies a lot over the time constant of a power transformer.
A transformer may take an hour to get to final temperature at a given power level. Anything that varies in that hour goes into the average. Typical music on radio stations and such has a crest factor of about 20 to 1. That is, the peak power is about twenty times the average. So mostly you get away with transformer VA ratings equal to or perhaps half of the rated RMS power of the amp. VA ratings of 100% of the power amp rating are quite conservative. VA ratings bigger than the RMS power of the amp are overkill by any yardstick I've ever seen. And those are the hifi yardsticks, where power supply sag is treated like an unmentionable social disease.
The thing is you almost can't prove than any transformer that doesn't go into meltdown isn't a good answer. Alternatively, there are many acceptable answers. It all depends on what you take as starting assumptions."
R.G.
____________________________________________________________________________________________________________
"The (perfed) chipamp is now getting hot at idle now, so I'm afraid I've got some oscillation going on. I'm wondering(out loud) if I need to add a Zobel to the output and what values to use. I've been reading in the DIY chipamp sites about "snubberizing", but I can't say it's making any sense to me.
This all may just mean that the heatsink(being too small and fan not able to keep up) issue is what is going on. I'm waiting for one to come in and then we'll see. Hopefully with a proper heatsink, it'll idle cool at least.
I just never had the chip getting hot at idle or low volumes before."
RDV

"Yes, it seem it might be self oscillating. Do you have the load (speaker) connected? These things don't like to be unloaded. Probably if you put a voltmeter set for AC voltage reading at the output you will be able to tell if the IC is actually oscillating.
The snubber thing is usually a small value resistor with a cap to ground loading the output, like 10 ohms 1/2W + 100nF or so. Check the datasheet to see if they suggest using something like this.
Also, the oscillation may be due to the layout and/or wiring around. Try rearranging wires, keeping input signals and control pots away from the output, etc.
Regards."
STM

"Every solid state power amp needs a Zobel network.
Every solid state power amp needs inductive catch diodes.
Every solid state power amp needs a damped series inductor.
Every solid state power amp needs power supply decoupling right at the circuit.
These are things, just like the wrapper on an effect, that are not particularly interesting or exciting - until it doesn't work.
stm's rignt - ten ohms and 0.1uF is the standard Zobel, and it seems to work for almost all speakers. And that wire arrangement is important.
However, he's incorrect about loading. Solid state amps and tube amps have opposite preferences about loading. Solid state amps work fine into arbitrarily high load resistances, including no load resistance connected at all. They tend to have trouble with too-low load resistances, including dying. Tube amps get unstable and oscillate, sometimes to death, when they are unloaded, but cope with overloads (too-low a load resistance, right down to dead shorts) just fine. They're different. However the conventional wisdom about musical instrument amps is not to unload them. It's a legacy of the tubes-only era."
R.G.

"I forgot to say that with prerecorded music(bypassing the TV Preamp/Limiter) this amp doesn't even get hot, barely warm even. With the guitar amp thing going it heats up almost instantly. I wonder if I should cap couple the input of the poweramp section?"
RDV

"It's always a good idea to cap couple if you can to prevent rogue DC from getting in.
I'm not too surprised that it doesn't warm up on prerecorded music. There's a generally accepted rule that prerecorded music has a crest factor (peak power to average power) of about 20db. The peaks are 20db larger than the average level. So if you play it cleanly, the average level barely warms the heatsinks. Not so for guitar by itself, especially with a square wave fuzz going into it. The worst thing you can to to your amp is not to play full-bore sine waves or music through it. The worst thing you can do is to play half-maximum swing square waves. That means that each output device is alternately handling half the power supply voltage and half the full current. That's the peak power dissipation point in the output transistors. Playing louder square waves actually lets them cool off some from there.
... of course, you could have some DC leaking in too..."
R.G.

"I've never given up on the original chipamp I perfed in the beginning of this. It was heating up and shutting off no matter how much heatsink or what size load I put on it.
The last couple of things I tried seemed to do the trick. I replaced the polys(.1µF) on the PS bypass with monolithic ceramic and moved them right by the chip. I changed the zobel to a 2.7 ohm and another monolithic ceramic. I had removed the 10k pot early on so in case that was the problem I installed a 22k from input to ground. The feedback resistor it seems was hanging by a thread and broke off when I touched it. EEK. Replaced that also and fired it up without the preamp and it didn't get hot and sounded okay too. Cool. I've spent about 2 weeks trying to figure this out."
RDV
_________________________________________________________________________________________________________
"Did you build your lm3886 to match the GGG schematic or did you just do it exactly the way the chipamp.com stuff is set up?"
Joecool85

"I followed the GGG schem, then started experimenting using a bunch of various schems as guides. The way my perfed one stands right now is kind of a combo of GGG's and chipamp's along with a PS borrowed from Rod Elliott. A good place to look as they have a chipamp forum is www.DIYAudio.com"
RDV
_________________________________________________________________________________________________________

2 ea. LM3886TF - $13.00
1 ea. Dual Mono LM3886 PCB from www.chipamp.com - $26.00*w/shipping
1 ea AC Receptacle - $.50
1 ea Computer AC Cable - $0.00
1 ea. AVEL Y236503 160VA 22V+22V TOROIDAL TRANSFORMER - $46.99 w/shipping
1 ea Chassis 2 rack space aluminum enclosure - $0.00**
2 ea 10,000µF 50v Electrolytic Capacitor - $7.34
6 ea 100µF 50v Electrolytic Capacitor - $.60
2 ea 47µF 50v Electrolytic Capacitor - $.14
1 ea 10µF 50v Electrolytic Capacitor - $.07
6 ea .1µF 50v Poly Capacitor -  $1.44
8 ea 8.0 Amp 600 Volt Diodes - $10.48
1 ea 1N4004 Diode - $.05
3 ea 10k Metal Film Resistor - $.30
4 ea 22k Metal Film Resistor - $.40
4 ea 1k Metal Film Resistor - $.40
2 ea 680ohm Metal Film Resistor - $.20
2 ea 2.7ohm 2watt Metal Film Resistor - $.64
2 ea 2.2ohm 2watt Metal Film Resistor - $.64
1 ea Green LED - $0.00***

Total: $109.19

*Came with 2 PS PCBs, only used 1, saved 2nd for next project.
**Broken(and unfixable IMO) Carver CD Player from the 80's had 2 space rack mount chassis. A comparable chassis from Par-Metal.com would be around $50.00
***Freebie from the wonderful Brian Marshall. LEDs are cheap.
RDV

"That's a good illustration of something I keep saying - a power amp is mostly power supply and other things that are NOT the power amp itself.
Here's what I think we can credit to the power amp circuit:

2 ea. LM3886TF - $13.00
1 ea. Dual Mono LM3886 PCB from chipamp.com - $26.00*w/shipping
6 ea 100µF 50v Electrolytic Capacitor - $.60
2 ea 47µF 50v Electrolytic Capacitor - $.14
1 ea 10µF 50v Electrolytic Capacitor - $.07
6 ea .1µF 50v Poly Capacitor -  $1.44
1 ea 1N4004 Diode - $.05
3 ea 10k Metal Film Resistor - $.30
4 ea 22k Metal Film Resistor - $.40
4 ea 1k Metal Film Resistor - $.40
2 ea 680ohm Metal Film Resistor - $.20
2 ea 2.7ohm 2watt Metal Film Resistor - $.64
2 ea 2.2ohm 2watt Metal Film Resistor - $.64
Which adds up to $43.88 if I added right.

And here's the non-power-amp-circuit stuff:

1 ea AC Receptacle - $.50
1 ea. AVEL Y236503 160VA 22V+22V TOROIDAL TRANSFORMER - $46.99 w/shipping
2 ea 10,000µF 50v Electrolytic Capacitor - $7.34
8 ea 8.0 Amp 600 Volt Diodes - $10.48
1 ea Chassis 2 rack space aluminum enclosure - $0.00**
1 ea Computer AC Cable - $0.00
1 ea Green LED - $0.00***
Which adds to $65.31.

The power transformer itself cost more than all the power amp circuitry and purchased PCBs.

The imbalance gets even worse when we notice that all the free stuff was non-poweramp, and that the cabinet could have added more cost than the power amp circuitry; even worse when we note that the PCBs could have been made at home for under $10.

This is the kind of stuff that people forget when they decide to build a big power amp.

By the way - you did a great job of holding costs down. Getting to under $1 per watt is a tricky goal in any kind of home built power amp that looks presentable."
R.G.
------------------------------------------------------
I hope this is of some help to someone.

RDV