Welcome to Solid State Guitar Amp Forum | DIY Guitar Amplifiers. Please login or sign up.

March 29, 2024, 08:10:02 AM

Login with username, password and session length

Recent Posts

 

Chip or Discrete PwrAmp-rebuild

Started by phatt, December 17, 2013, 07:47:35 AM

Previous topic - Next topic

phatt


I recently purchased some kit amps (LM3886) But seems these have trouble driving 4 Ohms,, hum got me thinking again maybe I should just build a simple discrete circuit and be done with it.
My guess is with 35~37 volt rails and 4 x TO3 3055/2955 devices I'd have little trouble driving an extra speaker when needed.

I'm not after the Big wattage,, just the load drive ability.

Here is one that looks like a good candidate.
http://sound.westhost.com/project03.htm

With 4 power transistors it should drive low loads without much issue.
BUT I want to keep the bandwidth similar to the Laney (KD50) which I note has a hi-pass on the input. (C2 on schematic)

As I use this for both keys and guitar,,I don't want 10~30 Khz bandwidth as digital synths can get gawd damned awful harsh if played through full bandwidth equipment. :grr

So Q is; Can I tweak the Project 03 or similar circuit to react like my Laney KD50 pwramp?
Maybe the diff is negligible and hardly worth the effort but I know from experimentation with other gear there is sometimes a notable difference.
Some power amps have heaps of bottom end yet others do not.
Depending what you are chasing sometimes this is good but maybe not so good if you have a digital keyboard with massive bass output and patches can be quite different, can be a ripe royal pain
trying to balance your sound patches if there is a bottom heavy or top end peaky poweramp.

I know from earlier attempts and simulations that the value of R6 (820R on Laney schematic) does affect freq somewhat but this can upset other parameters,, even going up in smoke.

The other thing of interest is the heat from the laney is not extreme even though the heat sink is about as cheap as you can get the back panel never gets hot, just warm and I do belt this thing hard at gigs so it obviously has no issues with heat.
I don't want to build something that runs on the bleeding edge and needs a fridge mechanic to keep it cool.

I've got plenty of PSU bits to work with just need to drill the TO3 holes in alloy to mount the 4 Pwr Tr's and some bread-boarding will begin.
Looking for clues guys,, open for suggestions.

I could just rebuild the Laney design but after reading Rods insightful observations it might be better to setup a bread-board test and see what I can discover.
I like how this page explains the difference between output designs.
http://sound.westhost.com/articles/cmpd-vs-darl.htm#s5

Adding the laney PDF schemo for reference but it's the KB80 but they are very similar.
I've included another schem of The actual values on my KD50 Amp.
My guess is the KB80 runs higher rail voltages? Mine reads +/-37VDC.

Obvious differences are;
R1 =6k8 not 1k
C2 =4n7 not 2n2
D1 and D2 are inverted
Tr5 and TR6 are different

Also C1 is now 47uF which is my attempt to get a little more low end response.
Oh yeah I don't need the protection stuff as I've had wonderful success with those
PTC units (poly thermal fuses) also makes less tracks on the PCB.

Phil.

joecool85

It has always been my understanding that LM3886 is fine driving 4 ohms so long as you have a large enough heatsink.
Life is what you make it.
Still rockin' the Dean Markley K-20X
thatraymond.com

phatt

Hi Joe,
Yeah you do need to read the fine print as I found out. :'(

I posted this a while back.
http://www.ssguitar.com/index.php?topic=3105.0

And just to repeat what was written in the corner of the label,


"Without looking hard enough I purchased 2 kits which I assumed where LM3886 powered BUT are instead LM3876 chips and did not notice till after sale that there maybe issues with 4 Ohm loads.
A note on the paper work clearly states that IF you wish to use 4 Ohm speakers you need to use LM3886 instead. Also adding that even with the LM3886 to keep supply rails below 30 VDC."

Bum,,So that plan came to a grinding halt.

Anyway,, Not much in the way of discrete pwramp design from ground up so I might inspire others to have a go at building something like this. :tu:
Phil.

Enzo

Never lose sight of the fact that the output power ALL comes from your power supply, not the output transistors.  Many circuits that "have trouble" driving a load are really just underpowered.   The output transistors or IC are nothing more than control valves for the power supply.

So when you say a chip amp cannot drive a 4 ohm load well, just what symptoms do you observe when it tries?

You mention four output transistors, but the circuit you posted has only two output transistors.   he does mention adding a couple near the end.   But you need to propose a power supply before this makes any sense.  A proper transformer that will provide sufficient current being the central element.

I don't know what amps you have tried, but in my experience, power amps all tend to be wide range circuits, it is the speaker more than anything that limits bandwidth, especially so in a guitar amp.  The enclosure being what makes one amp boomier than another.   If you find synths sound harsh, could it be the amps in question are playing through speakers with piezo tweeters?

In any case, I don;t expect tone shaping in a power amp.   C2 on your Laney KD50 is not a high pass circuit, in fact it is a low pass.  A shunt capacitor like that rolls off the high end.   In general, a cap like that is not there for tone, it is there to keep REALLY high frequencies out of the amp.  Like above audio.  No need for the amp to amplify radio signals - which it otherwise could do.

The job of the preamp is setting up the tone you want.    COnsider the venerable Peavey 400BH power amp.  A proven reliable design, this amp  was used in PA mixers, bass amps, guitar amps, keyboard amps, you name it.  Only differences were the preamps they put in front of it.


When comparing amps, it is important to know what you are looking at.  For example the D1,D2.   Inverted?  Sure, but all they are is a pair of 3.9v zeners wired back to back, so any signal peaks that exceed about 4 and a half volts will be clamped off.   it doesn't matter what order they are wired.

phatt

Hi Enzo,
Yes sorry I was not clear, there will be a *much larger Transformer* to supply this new setup. (same voltage, larger VA Rating)

OK forget speakers and other stuff, how about I start with the just the Amp.

As this thing is getting old and green stuff growing in places it might be wise to
build a new power supply and power amp stage anyway so why not make it a bit better.
Unless I'm sadly mistaken if I was to replace the power stage of my Laney it would make sense to run a quad of output devices as they share the load better.

I just want to keep the same response curve as some Poweramps go way down to DC and
I want to avoid that.
I recently hooked up one of my preamps to a brand new power amp and the bottom end
was was much greater than through my Laney. (Same speaker was used) So hence I'm
reluctant to just buy a kit and hope for the best.
Yes I'm very aware that Speakers count for a lot but IME power stage response comes
into this as well.

Now back to design;
As the laney uses same NPN output pair I could just double those for more current but I noted that Rod Elliot makes the point in his writings that using NPN devices for both sides of output stage is false economy these days and far better to implement both P and N output pair.

As PNP devices where more expensive to make in 1980 a lot of designs used only NPN pairs.
I also notice that Chip amps (including LM3886) are made to a budget and they also impliment NPN output pairs.

I get the distinct impression that Rod has done a fair bit of homework and real world testing so no point in me smoking another box full of perfectly good components just to find out same.

C2 is there to limit/cut higher frequencies, forgive my technical dyslexia.
Low pass filter = High cut.  High pass filter = Low cut.
For some confounding reason I always have trouble with those. :-[

Regards your Q; *So when you say a chip amp cannot drive a 4 ohm load well, just
what symptoms do you observe when it tries?*

I give picture as example.
I'm not a Teck by any stretch,, but I've been working with big welders and like equipment most of my working life and I know all about heat transfer,, well enough to know that if you are trying to pass a lot of heat through a narrow channel of material it heats up much faster than if the channel is broad.

So without even knowing anything about Amps or Watts it's rather obvious that those
two TO3 cases stand a much better chance of transferring heat from the case to surrounding heat sinks than the LM3886 beside it.

With chip amps the whole pre-drive circuit has to run at the same temperature as the
output devices and surely that is not an ideal situation?

As to D1 and D2 yep thanks for the clarification,, I guessed it made no difference but better to ask when not sure.

Thanks for the input, Phil.

Enzo

OK, I might be all wet, this is only an opinion, but I think you are rationalizing some.   There are different things in different contexts.  yes, two transistors can handle more current than one, but the question is whether the one is sufficient by itself.    A 4400 horsepower motor can handle pulling a trailer better than a 200 horsepower motor, but when all you are pulling is a lawn tractor, the extra capacity isn't important.

Same thing with the LM3886.   it iosn;t a question of whether larger hunks of metal can dissipate heat better, it is a matter of whether the chip is designed to cool itself sufficiently to meet its specs.  CHip amps are just everywhere, and they work well and reliably.  The whole hunk of silicon inside one may heat up to a certain level, but it is designed to work that way.  The heat transfer out of the thing via the mounting tab is trying to cool the hot spots on the silicon where the output transistors are.

COmplementary outputs versus quasi-complementary:  At some point they may have designed quasi amps because the PNP side of the equation was either more expensive or less reliable, and that is no longer the case.  But by the same token, the quasi circuit means you only have to stock one type of output transistor instead of two, and the circuit works well and reliably.  SO when I see a Peavey 400BH, a quasi amp circuit, I have no reason at all to consider redesigning it for true complementary.  It works great as it is.   And I can stuff MJ15003s into it any time without having to find some Mj15004s to go along.

National designed the LM3886 to be a reliable and simple to use power amp on a chip.  The fact is uses all NPN outputs inside is probably more about the fact it is easier to make all one type on the silicon than to make complements.  The chip amps are not just tiny parts on a board, it is one hunk of silicon.   Imagine if you wanted your computer printer to put a large black "X" on a white page.  NO problem.  Now imagine you wanted it to print a large white "X" on a black page.  You COULD get white toner and load black paper into the printer.  Or you could simply accept that it is easier to design all the images the same, and print black all around the area that remains white.   The design of the chip is sorta like that.

So the example of larger things cooling better doesn't really answer the question of how these smaller amps fall short when trying to drive a 4 ohm load.  In other words, I want to make sure I HAVE a problem before I start to cure it.


SO I guess my message might boil down to using a proven design.  All the baggage about why someone might use quasi versus complementary becomes moot.   ANy reliable circuit will still be reliable if you clone it.

J M Fahey

I would leave the Laney as-is, build an extra powered cabinet, same size as Laney or slightly larger so the KB50 sits on top easily and stuff the new cabinet with any speaker/power amp combination you wish.
And drive it with Laney line out.
Add it if not available.
Best of both worlds: the Laney on its own in smaller gigs or at home/rehearsal/friends' BBQ , etc. ; both together for big power/projection.

Roly

Quote from: phatt4 x TO3 3055/2955

The gain or hfe of a 2N3055 drops off as the collector current rises.  In spite of the "15Amp" Icmax rating in practical terms this means that a single pair starts to run out of puff with an 8 ohm load around 50Wrms.  I would consider two pair to be required for a 4 ohm load.
If you say theory and practice don't agree you haven't applied enough theory.

Enzo

Or drop a more modern transistor in their place, like MJ21194 and MJ21193, or even MJ15003 and MJ15004.

Roly

Yup, and not least for the astonishingly low fT of '3055's.  About the only reason to use them in fact is that they are almost dropping out of trees for free; I've seen them as low as 20 cents ea.
If you say theory and practice don't agree you haven't applied enough theory.

phatt

Thanks Enzo,
                  Yes I see what you say but can you appreciate that when you have a written warning (from a fairly reputable kit seller) about 4 Ohm loads on LM3886 chip Amp You don't just bang it all together and hope for the best.

Why would they give a warning if it's not an issue,, Head scratch?

Oh and I realize I did not fully Answer your Q about LM3886 symptoms.
Well I did build both those kits with LM3886 chips and ran them both from a 28-0-28 VDC with same signal at same time. One kit driving a 4 ohm speaker and the other driving an 8 Ohm speaker. (both 12 inch drivers of similar wattage and response)

Even at the lower recommended supply voltage the 4 Ohm distorted early on bigger signals. At full tilt they both distorted which was to be expected.
After 10 minutes the 4 Ohm kit was definitely running hotter to the touch.

Anyway,,,
I'll likely take *Mr Faheys* advice for now but with a twist.  ;)
The back panel of the Laney which is just a flat 3mm panel alloy extrusion mounts all the PSU and power amp board which detaches via 4 screws.

So If I recreate that back panel and make another PSu poweramp of a larger design then use it as another poweramp for now,,, then it will be simple to replace the Laney back panel if The Laney ever dies.

The Zener diodes for preamp supply blew about 6 months back So I'm concerned that it's days are numbered but it's my main Amp and gives me everything I need.
I have backup gear but it's more cumbersome and has no reverb. :'(

@ Roly, Yep thanks Mate, With amplifiers of the Laney era I've always gone by the rule of thumb that if you see 4 x TO3 devices on the back and the rails are around the 30~40 Volt mark then a fair bet it can drive 4 Ohm better than with only 2 TO3.

Well anyway,, I've drilled the mounting for 4 TO3 on a 5mm extrusion and in time will experiment with that and see if I can keep the smoke inside components.
Hey I might even have another poke around with Current feedback ideas.
Yep might even look at MJ15003/4 in time but 3055 are dirt cheap to test out ideas,,
I have a history of blowing up power transistors while trying out crazy ideas.  :-X
Thanks for the input chaps, Phil.