| Dominick22 |
I am paralleling 2 LM4780 chips per channel fo my amp.
I have the power wired, but I am unsure of how to wire the signal paths. Both IN AND OUT.
I have ideas but thought I better ask before I see that magic black smoke!!LOL
Do I just run the + signal to one chip and the - to the other and then the corresponding polarity to the output jack?
*A diagram would be nice*
I am using the kits from audiosector.com.
Thanks a ton,
Dominick |
|
|
| amt |
How about the schematic thats on the website, showing the connections?
amt |
|
|
| Bazukaz |
Read LM4780's datasheet.There is a schematic of paralell config.
If the power supply is wired correctly,you shouldn't smoke LMs that easy.
A common problem is when there is no insulation washer between chip and heat sink(chip has -Vs connected to its case). If the sink is grounded , the power supply is usually powerful enough to blow IC in chunks.I have done it by myself once :).
Regards ,
Lukas. |
|
|
| Dominick22 |
I dont think that my question was understood.
I have 2 chips already wired in parallel. Now I need to know how to parallel the paralelled chips!!
Dominick |
|
|
| joecool85 |
| If I'm correct, that would give you support of down to 1ohm...any reason to need it that low? |
|
|
| Dominick22 |
Good observation.
That is exactly why I am doing it.
I intend on using this amp on my Martin Logan Ascent i speakers.
I called the company direct and they said that the speakers run at 2.5 ohms @ 20k.
Dominick |
|
|
| joecool85 |
| 2.5 ohms...that would run fine just with a parallel lm4780. 1 x 4780 parallel should give you about 120w @ 2 ohms I'm pretty sure. |
|
|
| Dominick22 |
I already tried it and blew one of the chips in less than a minute!
Dominick |
|
|
| joecool85 |
Did you have a heatsink on it? It shouldn't be a problem at all. People use the paralleled lm4780 setup from brian at chipamp.com for 2 ohms.
**edit**
Here is the schematic they use:
http://chipamp.com/lm4780.pdf |
|
|
| Bazukaz |
LM3886s can stand continous short circuit and has lots of protection.
Adding more in parallel may even worsen results , if ICs have DC offsets greater than a few mV. |
|
|
| merlinx76 |
| I found it necessary to add the optional Ci capacitors (shown on the datasheet) for unity gain at DC on my parallel 4780. Without it the amp got quite warm before even hooking it up to speakers. Perhaps this is what happened to you? |
|
|
| Dominick22 |
I appreciate all of the suggestions, but I am sure that the amop cannot handle the ML speakers with one ship per channel. Before running the ML's, I used the amp for about 5 hours STRAIGHT on an 8 ohm load. Then, the next day, it blew with the logans in 1 minute.
I have since ONLY replaced the 1 chip and have probably used the amp on an 8 ohm load for about 50 more hours with no problems what so ever.
I already have the 2 chips per channel comleted....
I just need to no how to run the input/output signals.
UPDATE:
Last night I ran the sig in and sig g to both chips and ran the out from one chip and the out g from the other.
Funny thing is I could here a light distorted sound from the 8 ohm speaker.
When I turned the amp on, it was blaring so loud that I couldnt tell if the distortion was from the amp or if it was just too loud for the speaker.
I rolled the pot to the other side and the same thing occurs although this time I didnt get any light distorted sound prior to kicking on the amp???
Obviously I am doing something wrong but I think I am close.
HELP ME PLEASE,
Dominick |
|
|
| Bazukaz |
I think that the main problem is NOT related with low drive capability.
2.5 ohms is not so low.
I have had some experience with shorting outputs of LMs , and they stand this easy , even in bridged configs.
In your place , i would try just a single LM3886 and see what happens.
With paralelling , the chips may start to fight in some cases(especially if they oscillate).
Oscillation could be your problem.
Try placing ~0.5 ohm resistor at ouput of your amp.
Also , check if chips don't get hot while not playing sound.
Have you wired everyting to a single ground poit ? |
|
|
| Dominick22 |
I did notice that one of the two chips didn't get warm when I turned the amp on for a few secs.
But I still got a loud distorted sound.
Does this mean that one chip is not turning on?
I don't know what this means?
Dominick |
|
|
| Dominick22 |
I only need to use the out on one chip and the out ground on the other, right?
or,
Do I need to connect the out and ou ground of both chips to the speaker terminals?
Dominick |
|
|
| Bazukaz |
You connect out and out , ground to ground.
Thats paralelling. |
|
|
| Peter Daniel |
In order to parallel two parallel configured LM4780 chips, the best way it to position the boards back to back. This will simplify ground connections.
For both boards, you connect together: OG, SG, //OUT and IN. It's better to use two R4 resistors (one per board), however they will be in parallel and this will lower input impedance to 10K (this will also produce much less DC offset). Alternatively, you can substitude 22K resistors with 40K resistors, for 20K total shunt input impedance.
Everything else is configured the same as per schematic for parallel setup here: http://www.audiosector.com/lm4780%20amp.pdf
PS wires cannot be connected with short jumpers (between two boards) as their access pads are not symetrical , you need to use longer pieces of wire. Both grounds connections (for two boards) can be as short as space between the boards ( as the pads are symetrical here). |
|
|
| amt |
Sorry I didnt understand what you were doing originally.
If one chip isnt functioning, Im sure you are going to have problems. I researched paralleling four LM3886 a while back and was told that the resistors used on the outputs should be 1% or better. Im guessing that the ones used on Peters boards (R8,9) are such. It was also suggested that I use 4700uf caps on each output, after the resistors, and then combine the outputs. This is suppose to prevent any DC interaction between the amplifiers.
It appears that this may be what was done on the Meier power amp:
http://www.meier-audio.homepage.t-online.de/index.htm
I havent gotten mine completed, so I cant comment on the success of this arrangement. Hope this is of some help.
amt |
|
|
| Dominick22 |
I just tested each board seperatly and they are producing an audio signal to the speaker although with a very slight hum.
The wierd thing is that on one board, the chip never even warms up??
But both boards play?? The other chip gets warm regardless of the board with the audi signal??
Wierd huh?
IDEAS?
Dominick
Peter, I will try to understand you post!lol And execute! |
|
|
| Peter Daniel |
In combining 2 parallel LM4780 chips (4 x LM3886), it may be a good idea to increase output resistors to 0.22R or even a bit more.
I also have some chips that are measured for DC offset. Paralleling chips with similar DC offset could be advantegous here. |
|
|
| Bazukaz |
Hum may be caused by poor PCB layout(especially grounding).
What schematic are you using ? PCB layout ? Could you post it ? |
|
|
| jackinnj |
| quote: | Originally posted by Peter Daniel
In combining 2 parallel LM4780 chips (4 x LM3886), it may be a good idea to increase output resistors to 0.22R or even a bit more. |
Why?
Probably a better idea to match the gain resistors until the offset is nulled out.
the R*I^2 losses get pretty big when you double or quadruple the resistance values. why burn power when you can listen to it? |
|
|
| Peter Daniel |
I was using 0.22R resisitors myself and didn't notice any drawbacks.
What do you mean by matching gain resistors to null the offset? |
|
|
| jackinnj |
| should have said "matching gain and input resistors" to null the offset. |
|
|
| Dominick22 |
Bazukaz,
I bought 4 LM4780 chips complete with kits from Peter Daniel at audiosector.com.
I reccomend them 2-fold.
1st, I have built 1 amp already using his kits and I am more than pleased.
Second, as you can see, Peter is more than willing to offer his expertise...even if some of the things he says are above my skill level!!
So, go to audiosector.com and click on LM4780 Dual mono kit.
That is what I am working with.
TO EVERYONE: Thank you all for the replies although I am having some trouble understanding some of the terminology.
Maybe I will post some pics of my project in development!?
Dominick |
|
|
| Bazukaz |
Well , the kit looks good.It has support of paralell configuration,so values should be already matched and you should not need to do anything.
My advice(already mentioned) : add a resistor between your amp and speakers.you don't need large - ~0.5-1 ohm.It should be non-inductive.
Plug in and watch if your amp smokes :)
Now about instabilities : imagine one amp starts oscillate , at a frequency of few MHZ or so. Another will then try to compensate.
So both amps will be stressed in parralel config.Adding two more amps in paralell may even worsen the problem ... |
|
|
| Dominick22 |
Do I put tha .5/1 ohm reisitor on both the negative and positive leads or just one of them?
And which one would be the safer to try first?
Dominick |
|
|
| ChristianThomas |
Hi
You're driving electrostatics which get quite capacitative at HF. This could easily be the cause of instability, especially if layouty isn't ideal.
The most usual solution for this is an inductor on the output - Ie. after your R22. Get a 2W 2R2 resistors and wrap 8 to 12 turns of 0.7mm enamelled wire around it so the two are in parallel. This should just about eliminate the idea of the speaker being the source of your trouble - and you should be able to use just two chips in parallel.
Hope this helps
Christian |
|
|
| Bazukaz |
You just put 0.5 ohms resistor (non-inductive , 5 or more watts preferred) in series with loadspeakers.
Another way is to use L-R filter.
for this , you need a coil of around 1uH(15 turns on 1.5 cm diamter with 1.5 mm enamel wire).The inductor is then connected in paralell with 10 ohm 2 watt resistor).The filter is connected in series with loadspeaker. |
|
|
| Dominick22 |
UPDATE:
I want to thank everyone who is following this thread.
After some trial and error testing, I can now control the speaker volume with the pot however the audio signal is masked greatly by static. I can here music behind the static.
Also, BOTH chips are now heating up very quickly...even with a sink attached.
Baz, I tried using the resistor tip at the output and it only seemed to completely block the entire signal. Everything was reduced in violume wholly. It did not improve the sound.
I was looking at the original chip schematics from National and I saw the ci1 and ci2 for parallel config. I am unsure were to place this in the circuit??? Should I literally just create a lead off of 1 of the 4 V- pins??? Should I do 1,2, or all 4 V- pins???
I am reaching for ideas here. Maybe I shouldnt jump from a pin at all???
At least we are making progress.
Dominick |
|
|
|
