diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Chip Amps (http://www.diyaudio.com/forums/chip-amps/)
-   -   Building a Gainclone with LM3886 General Direction and PSU (http://www.diyaudio.com/forums/chip-amps/145035-building-gainclone-lm3886-general-direction-psu.html)

bmwpowere36m3 30th May 2009 04:56 PM

Building a Gainclone with LM3886 General Direction and PSU
 
Looking towards another project this time a speaker amp (stereo), expanding my horizons. This will be my first speaker amp, however have already conquered a CMoY, "SS" Millet, SOHAII, and a ϒ1.

There is so much info out their on this chip-amp its hard to make choices out of all the info :confused:, but also a majority of it is on the LM3875. So to start off, I want to know some basic info...

Looking at the spec-sheet (LM3886) a Vcc of 30V seems like a good compromise between running a 4 or 8 ohm load. At this point I'm not sure what speakers I want to run, but I want keep this as universal as possible and powerful enough (just in case :veryevil:). At 30V, 4-ohms, the output is ~85W and at 8-ohms ~45W. I'm not looking for SPL > SQ, more interested in SQ > SPL.

However looking at some of the distortion figures, I guess what's the problems with running too high a Vcc, heat dissipation of the chip and distortion? What's a good compromise? Is it worthwhile to build a PSU, that can output 30-40V?

Next up the trafo, how big (VA)? The chip specs a max output of 135W, so a stereo setup (2x chips) would need a trafo to support a max of 270VA.... correct or too much? I see other builds specing anywhere from 160-300VA.

Finally looking at two PSU designs to choose from, the first part of chipamp.com kit (CarlosFM):

http://www.chipamp.com/images/ps.gif

A regular "snubberized PSU", they mention with a 18V trafo the output is ~25V and with a 22V trafo is ~34V. Finally the VA rating should be 220-330VA.

Another is PSU is CarlosFM's Regulated Snubberized PSU:

http://myweb.tiscali.co.uk/nuukspot/...ub.reg.psu.png

Again the same basic principle, but I guess a better design as the output is now more "regulated" via the LM338. From what I read over on the site which hosts this PSU design, it did improve the sound quality of the gainclone.

Finally this PSU, which appears to be a variation of the above one (CarlosFM's Regulated Snubberized PSU), with more capacitance (37,6k uF vs. 9,4k uF):

http://feuerbacher.net/DogBreath/Chi.../UnregPart.jpg

http://feuerbacher.net/DogBreath/Chi...338PSU/Reg.jpg

So is it worth building, I'm kinda leaning towards it and it doesn't seem anymore difficult to build. Here he lists to use a 2x 30V 300VA trafo, so again is 250-300VA enough?

Are the secondary voltages & VA they list for both these designs the same specs as called out by manufacturers like Avel Lindberg? I remember when building my SOHAII there was talk about the listed specs were for unloaded trafo's vs. loaded.

pacificblue 30th May 2009 07:05 PM

Re: Building a Gainclone with LM3886 General Direction and PSU
 
Quote:

Originally posted by bmwpowere36m3
However looking at some of the distortion figures, I guess what's the problems with running too high a Vcc, heat dissipation of the chip and distortion? What's a good compromise? Is it worthwhile to build a PSU, that can output 30-40V?
The LM3886 is specified for 28 V supply, when used with 4 Ohm loads and 35 V supply, when used with 8 Ohm loads. Those figures are valid, when the amplifier is loaded. At idle the voltages may be higher. An unregulated supply will sag, when loaded. You can take advantage of that, as shown in the BPA-200 application note. There a 60 V CT transformer is used with an 8 Ohm load.
You should also be aware that those figures are for the unisolated package. The isolated package forces lower ratings due to worse heat dissipation.

Quote:

Originally posted by bmwpowere36m3
Next up the trafo, how big (VA)? The chip specs a max output of 135W, so a stereo setup (2x chips) would need a trafo to support a max of 270VA.... correct or too much? I see other builds specing anywhere from 160-300VA.
Any rating that is bigger than the combined output power of the amplifiers should do.

Quote:

Originally posted by bmwpowere36m3
Finally looking at two PSU designs to choose from
The general tendency is that unregulated PSUs sound more dynamic, while regulated PSUs sound more detailed.
Why don't you start with an unregulated PSU and try, if you like the sound? You can add regulation later.
Nuuk has tried both. Maybe his verdict helps you find a decision.

Quote:

Originally posted by bmwpowere36m3
Are the secondary voltages & VA they list for both these designs the same specs as called out by manufacturers like
Transformer voltages are always given at nominal output. At idle they are higher by a certain percentage that is given as regulation.
Rail voltages are given as ((nominal transformer voltage X 1,41) - voltage drop across rectifiers). They sag, when the power supply is loaded. Rule of thumb is 10 %. Some designers rely on that sag for overload protection.

A well designed regulated supply will not sag, but it needs a higher transformer voltage, so that the voltage before the regulator is always a few volts higher than the output voltage. That means it also needs a higher power rating, because the voltage difference between in- and output times the current is converted into heat and reduces the power that is available for the amplifier by that energy. Therefore you also need additional heatsinking for the regulators. All that together makes for a significantly higher price and bigger space requirements.

bmwpowere36m3 30th May 2009 11:40 PM

Re: Re: Building a Gainclone with LM3886 General Direction and PSU
 
Quote:

Originally posted by pacificblue

The LM3886 is specified for 28 V supply, when used with 4 Ohm loads and 35 V supply, when used with 8 Ohm loads. Those figures are valid, when the amplifier is loaded. At idle the voltages may be higher. An unregulated supply will sag, when loaded. You can take advantage of that, as shown in the BPA-200 application note. There a 60 V CT transformer is used with an 8 Ohm load.
You should also be aware that those figures are for the unisolated package. The isolated package forces lower ratings due to worse heat dissipation.


Any rating that is bigger than the combined output power of the amplifiers should do.


The general tendency is that unregulated PSUs sound more dynamic, while regulated PSUs sound more detailed.
Why don't you start with an unregulated PSU and try, if you like the sound? You can add regulation later.
Nuuk has tried both. Maybe his verdict helps you find a decision.


Transformer voltages are always given at nominal output. At idle they are higher by a certain percentage that is given as regulation.
Rail voltages are given as ((nominal transformer voltage X 1,41) - voltage drop across rectifiers). They sag, when the power supply is loaded. Rule of thumb is 10 %. Some designers rely on that sag for overload protection.

A well designed regulated supply will not sag, but it needs a higher transformer voltage, so that the voltage before the regulator is always a few volts higher than the output voltage. That means it also needs a higher power rating, because the voltage difference between in- and output times the current is converted into heat and reduces the power that is available for the amplifier by that energy. Therefore you also need additional heatsinking for the regulators. All that together makes for a significantly higher price and bigger space requirements.

Thanks, after reading Nuuk's "journay" is says the best PSU was a discrete PSU by Pedja (Module R). After reading its description I wonder if I could use it with what I want to do?

pacificblue 31st May 2009 07:36 AM

Yes, you can.

bmwpowere36m3 31st May 2009 08:39 PM

Quote:

Originally posted by pacificblue
Yes, you can.

After reading the description, 1 board can support 30 W into a 8 ohm load (2x stereo) ~2.7 A output with a 200 VA trafo. Two boards can support 5.4 A ? with a 500VA trafo. After doing some calculations last night, if I wanted to power the LM3886 to its maximum while properly heat sinking it.... the Vcc would be 30 V and with a 4 ohm load, output ~85 W continuous, Ioutput 6.5 A. So two chips in stereo, to operate @ 85 W continuous would require a PSU capable of delivering minimum 13 A, correct? At an 8-ohm load, output ~45 W continuous, Ioutput 3.4 A. So two chips in stereo, to operate @ 45 W continuous would require a PSU capable of delivering minimum 6.8 A, correct?

Am I looking for unrealistic demands on the PSU, since from what I can tell some PSU designs (CarlosFM) are limited to 5 A due to the LM338 and even then it would be at its limit and hot.

pacificblue 31st May 2009 09:26 PM

Quote:

Originally posted by bmwpowere36m3
After doing some calculations last night, if I wanted to power the LM3886 to its maximum while properly heat sinking it.... the Vcc would be 30 V and with a 4 ohm load, output ~85 W continuous, Ioutput 6.5 A. So two chips in stereo, to operate @ 85 W continuous would require a PSU capable of delivering minimum 13 A, correct? At an 8-ohm load, output ~45 W continuous, Ioutput 3.4 A. So two chips in stereo, to operate @ 45 W continuous would require a PSU capable of delivering minimum 6.8 A, correct?

Am I looking for unrealistic demands on the PSU, since from what I can tell some PSU designs (CarlosFM) are limited to 5 A due to the LM338 and even then it would be at its limit and hot.

More like looking for unrealistic demands of the LM3886.

Are you going to use those amplifiers to reproduce continuous sinus waves? With a normal music signal you won't need continuous output above a few Watts.

The LM3886 is specified as 68 W into 4 Ohm for several reasons. You should not try to squeeze more out of it.

For normal music program one PSU board will be sufficient for 2 channels. Heatsinks on the MUR860s won't hurt however.

bmwpowere36m3 1st June 2009 01:24 AM

Quote:

Originally posted by pacificblue

More like looking for unrealistic demands of the LM3886.

Are you going to use those amplifiers to reproduce continuous sinus waves? With a normal music signal you won't need continuous output above a few Watts.

The LM3886 is specified as 68 W into 4 Ohm for several reasons. You should not try to squeeze more out of it.

For normal music program one PSU board will be sufficient for 2 channels. Heatsinks on the MUR860s won't hurt however.


Ha ha, nope won't be playing sine waves. So in that case building it around 68W @ 4 Ω, I can go with a 25-30 V secondary trafo and a PSU with regulation to keep it anywhere from 25-30 V Vcc. The one PSU board you refer to is Pedja's Module R, correct. In which case he recommends a 200VA trafo. The reason I started thinking about it was that other designs and users have commented on 160VA trafo's being the minimum required and something around 250-300VA ideal. Would running that size trafo on one of Module R PSU's allow for more current than it can handle?

pacificblue 1st June 2009 07:16 AM

[QUOTE]Originally posted by bmwpowere36m3
So in that case building it around 68W @ 4 , I can go with a 25-30 V secondary trafo and a PSU with regulation to keep it anywhere from 25-30 V Vcc.[QUOTE]
Yes.

Quote:

Originally posted by bmwpowere36m3
users have commented on 160VA trafo's being the minimum required and something around 250-300VA ideal.
Those comments were probably refering to unregulated power supplies, and a lot of them rely on hearsay.
The transformer must be able to deliver the output power plus the losses in the circuit. Therefore a transformer should be at least as big as the output power.

With an unregulated power supply the losses are not too big, and there are only few and short moments, when the peak output power is actually delivered. So you could even get away with a transformer that is a bit smaller.
People like to oversize transformers to around three times that number, based on the theory that the transformer only recharges the capacitors for about 1/3 of the actual time, must therefore deliver three times the current and would work in its core saturation region then. Well-designed transformers that are designated as transformers for rectifier operation take that already into account and have a sufficient core diameter to not saturate on these conditions.
Another reason for oversizing is to keep regulation low. In the range from 160-300 V a drop from around 12% to around 8 % can be expected. No big deal.
The real reason, why oversized transformers can improve the sonic preformance is probably the temperature rise that makes saturation take place at a lower level. Transformers are usually specified for temperatures of 40-50 C. In an amplifier enclosure the temperatures can be higher and few people conciously derate their transformers according to the temperature. Oversizing them is more or less unconcious derating.

Circuit losses are considerable, if you use a regulated power supply, so a transformer for that task must be significantly bigger than the output power. Pedja Rogic's recommendation of 200 VA seems reasonable.

Quote:

Originally posted by bmwpowere36m3
Would running that size trafo on one of Module R PSU's allow for more current than it can handle?
The LM338's datasheet has some information on this. It says that it is "capable of supplying in excess of 5 A" and "allows current peaks of up to 12 A to be drawn for short periods of time. And it says that the LM338 comes with "thermal overload protection and safe area protection for the power transistor". There seems to be no need to worry about the LM338.

bmwpowere36m3 1st June 2009 04:13 PM

Quote:

Originally posted by pacificblue
The LM338's datasheet has some information on this. It says that it is "capable of supplying in excess of 5 A" and "allows current peaks of up to 12 A to be drawn for short periods of time. And it says that the LM338 comes with "thermal overload protection and safe area protection for the power transistor". There seems to be no need to worry about the LM338.
Well the Module R PSU doesn't use a LM338, here is it's schematic:

http://i61.photobucket.com/albums/h6...mblyManual.jpg

Either way, I found a link to the source's forum: http://www.diyhifi.org/forums/viewtopic.php?f=5&t=177 So you'll figure out the details of the PSU there. Thanks again.

bmwpowere36m3 5th June 2009 02:01 AM

Now switching gears to the amp-section itself, I've been contemplating using chipamp.com's PCB for my build. My question is the integration of my headphone amp (Millet "Starving Student" Hybrid) as a preamp into the power amp section. After reading the LM3886 data sheet the implementation of the input resistance is confusing me. It says that the feedback resistance should be matched with the input resistance, i.e. the same and that the ideal input resistance is 100k minimum. By setting the feedback resistance equal to 100k it minimizing DC offsets at the output. Is this correct? Finally as I understand it the purpose of Rin is to keep the chipamp stable when there is no "preamp" or source connected. So with one connected is it still necessary to worry about it?

Basically this is the headphone circuit used:

http://i61.photobucket.com/albums/h6...dSchematic.jpg

and this is Chipamp.com's LM3886 Kit:

http://i61.photobucket.com/albums/h6...s/Picture1.png

For one, I assume the output impedance of the millet is 2k, right as formed by the RC. Second I know the output capacitors are large in efforts to maintain bass when low-impedance headphones where plugged in. With the high impedance of the LM3386, they could be smaller yet preserve "bass" or bandwidth, correct?

What is the best way to couple these two sections, noting that the headphone amp will ONLY be used with the chipamp. Finally I read some suggestions as to various changes to the selected resistances and capacitors used in chipamp's PCB....


All times are GMT. The time now is 11:36 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2